Biological sequence extraction and retrieval methods, devices and electronic equipment

Abstract

This invention provides a method, apparatus, and electronic device for biological sequence extraction and retrieval. The method includes: acquiring an initial biological sequence from a biomedical document, and sequence attribute information associated with the initial biological sequence; performing sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result; and extracting the initial biological sequence whose attribute verification result is successful as a target biological sequence. The biological sequence extraction and retrieval method, apparatus, and electronic device provided by this invention can effectively improve the accuracy and efficiency of biological sequence identification and extraction.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, and electronic device for biological sequence extraction and retrieval. Background Technology

[0002] Life information is contained in gene sequences, which, through transcription and translation, synthesize proteins with specific amino acid sequences, thereby controlling the traits of organisms. Research and development in the biomedical field mainly involve the structure and function of biomolecules such as DNA, RNA, and proteins; therefore, the effective utilization of biological sequence information is crucial for the development of the biomedical field.

[0003] Biomedical documents contain a wealth of biological sequence information. Analyzing and extracting this information, and then integrating it to create a biological sequence database, is crucial for advancing biomedical and biological research. However, existing biological sequence identification and extraction technologies suffer from low accuracy, and the extracted sequences require manual calibration, which not only increases workload but is also highly inefficient. Summary of the Invention

[0004] This invention provides a method, apparatus, and electronic device for biological sequence extraction and retrieval, which solves the problems of low accuracy and low efficiency in biological sequence identification and extraction in the prior art.

[0005] This invention provides a method for biological sequence extraction, comprising:

[0006] Obtain the initial biological sequence from documents in the biomedical field, as well as the sequence attribute information associated with the initial biological sequence;

[0007] Based on the sequence attribute information, the initial biological sequence is subjected to sequence attribute verification to obtain the attribute verification result;

[0008] The initial biological sequence whose attribute verification result is successful is extracted and used as the target biological sequence.

[0009] According to a biological sequence extraction method provided by the present invention, the sequence attribute information includes a first sequence length and / or a first sequence type;

[0010] The step of verifying the sequence attributes of the initial biological sequence based on the sequence attribute information to obtain the attribute verification result includes:

[0011] The sequence length of the initial biological sequence is calculated as the second sequence length. Based on the second sequence length and the first sequence length, the initial biological sequence is validated for sequence length to obtain a length validation result; and / or, based on the first sequence type, the initial biological sequence is validated for sequence type to obtain a type validation result.

[0012] The attribute verification result is determined based on the length verification result and / or the type verification result.

[0013] According to a biological sequence extraction method provided by the present invention, the step of performing sequence type verification on the initial biological sequence based on the first sequence type to obtain a type verification result includes:

[0014] Determine the first unit type of the basic units constituting the first sequence type of biological sequence, and determine the second unit type of each basic unit contained in the initial biological sequence;

[0015] Based on the first unit type and the second unit type, the initial biological sequence is subjected to sequence type verification to obtain the type verification result.

[0016] According to a biological sequence extraction method provided by the present invention, determining the attribute verification result based on the length verification result and / or the type verification result includes:

[0017] If the length verification result is successful and the type verification result is unsuccessful, the initial biological sequence is corrected based on the correction method corresponding to the first sequence type, and the sequence attributes are verified based on the corrected initial biological sequence to obtain the corrected length verification result and type verification result.

[0018] If both the corrected length verification result and the type verification result are successful, the attribute verification result is determined to be successful.

[0019] According to a biological sequence extraction method provided by the present invention, the step of obtaining an initial biological sequence from a biomedical document and sequence attribute information associated with the initial biological sequence includes:

[0020] Obtain documents in the biomedical field;

[0021] When the document in the biomedical field includes images, text recognition is performed on the document to obtain multi-line recognized text;

[0022] The multi-line recognized text is classified to obtain the text type of each line of recognized text, and the initial biological sequence and sequence attribute information are determined from the multi-line recognized text based on the text type.

[0023] Based on the sequence identifier corresponding to the initial biological sequence and the sequence identifier corresponding to the sequence attribute information, the sequence attribute information associated with the initial biological sequence is determined.

[0024] According to a biological sequence extraction method provided by the present invention, the acquisition of biomedical documents includes:

[0025] Obtain the initial domain file;

[0026] If the initial domain file includes images, the images are classified into sequential pages to obtain the sequential page classification results of the images;

[0027] Based on the images that belong to the sequence page classification results and the readable text in the initial domain file, the biomedical domain file is determined.

[0028] This invention also provides a biological sequence retrieval method, comprising:

[0029] Obtain the target search terms;

[0030] The biological sequence library is used to determine the target biological sequence that matches the target search term. The biological sequence library is constructed based on the target biological sequence determined by any of the biological sequence extraction methods described above.

[0031] Display the target biological sequence that matches the target search term.

[0032] The present invention also provides a biological sequence extraction device, comprising:

[0033] The acquisition unit is used to acquire the initial biological sequence from a document in the biomedical field, as well as the sequence attribute information associated with the initial biological sequence.

[0034] The verification unit is used to verify the sequence attributes of the initial biological sequence based on the sequence attribute information and obtain the attribute verification result.

[0035] An extraction unit is used to extract the initial biological sequence whose attribute verification result is successful, as the target biological sequence.

[0036] The present invention also provides a biological sequence retrieval device, comprising:

[0037] The term acquisition unit is used to acquire target search terms;

[0038] A sequence retrieval unit is used to determine a target biological sequence that matches the target search term through a biological sequence library, wherein the biological sequence library is constructed based on the target biological sequence determined by any of the biological sequence extraction methods described above.

[0039] The sequence display unit is used to display target biological sequences that match the target search term.

[0040] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement any of the biological sequence extraction or retrieval methods described above.

[0041] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the biological sequence extraction or retrieval method as described above.

[0042] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the biological sequence extraction or retrieval method as described above.

[0043] This invention provides a biological sequence extraction and retrieval method, apparatus, electronic device, and storage medium. By acquiring initial biological sequences from biomedical documents and sequence attribute information associated with the initial biological sequences, the method verifies the sequence attributes of the initial biological sequences based on the sequence attribute information. This allows for the extraction of successfully verified initial biological sequences as the final target biological sequences, ensuring high accuracy in biological sequence identification and extraction. Furthermore, it eliminates the need for manual intervention and calibration, further improving the efficiency of biological sequence extraction. Compared with existing technologies, this invention determines the target biological sequence based on attribute verification results, resulting in higher accuracy and reliability of the extracted biological sequences. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0045] Figure 1 This is one of the flowcharts of the biological sequence extraction method provided by the present invention;

[0046] Figure 2 This is a schematic diagram of a portion of the biological sequences in the sequence listing of a patent document in the biomedical field provided by this invention;

[0047] Figure 3 This is the second flowchart of the biological sequence extraction method provided by the present invention;

[0048] Figure 4 This is a flowchart illustrating the biological sequence retrieval method provided by the present invention;

[0049] Figure 5 This is a schematic diagram of the biological sequence extraction device provided by the present invention;

[0050] Figure 6 This is a schematic diagram of the biological sequence retrieval device provided by the present invention;

[0051] Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0053] In the description of the various embodiments of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" and "second" may explicitly or implicitly include one or more of the stated features.

[0054] Biological sequences refer to the sequences of biomolecules such as DNA (Deoxyribonucleic acid), RNA (ribonucleic acid), or proteins, and are among the most fundamental data in biological research. With the development of high-throughput sequencing technology and large-scale bioinformatics databases, the amount of biological sequence data is constantly increasing, and it has become a major data source for bioinformatics research.

[0055] Currently, biopharmaceutical companies typically search for the biological sequences of similar biopharmaceuticals before screening or designing a new drug to determine if the proposed drug infringes on intellectual property rights. If a similar biological sequence has already been protected by another company, the company will make adjustments as early as possible, such as terminating drug development, to save on R&D costs. Therefore, analyzing biological sequence information in biomedical documents and integrating it to create a biological sequence database is crucial for advancing biomedical and biological research.

[0056] Biomedical documents contain a wealth of biological sequence information. Faced with this massive volume of documents, existing biological sequence identification and extraction technologies typically first use OCR (Optical Character Recognition) technology or related software to convert the documents into text format, then use regular expressions or other text processing techniques to filter and extract the biological sequence information. To ensure the accuracy of the extracted biological sequences, manual intervention and calibration are still required, resulting in very low efficiency. For documents with varying corpora and layouts, existing biological sequence identification and extraction technologies suffer from low accuracy. Furthermore, the documents often contain a lot of redundant information, and existing technologies require time-consuming and labor-intensive analysis of each element, further contributing to the low efficiency of biological sequence identification and extraction.

[0057] In response, this invention provides a biological sequence extraction method that automatically verifies the sequence attributes of the initial biological sequence by acquiring sequence attribute information associated with the initial biological sequence, thereby determining the target biological sequence based on the attribute verification results. This method can not only effectively improve the accuracy and efficiency of biological sequence identification and extraction, but also provide highly reliable underlying data support for applications such as genomics research, protein analysis, biopharmaceutical development, and food safety testing. For details, please refer to the following text.

[0058] Figure 1 This is one of the flowcharts of the biological sequence extraction method provided by the present invention, such as... Figure 1 As shown, taking the application of this method to a server as an example, the method includes:

[0059] Step 110: Obtain the initial biological sequence from the biomedical field documents, as well as the sequence attribute information associated with the initial biological sequence;

[0060] Specifically, biomedical documents refer to documents containing biological sequences in the biomedical field. These can include, for example, papers, journal articles, patent documents, etc., containing biological sequences. The file type of biomedical documents can be portable documents (PDF), images, etc., and this embodiment of the invention does not specifically limit this.

[0061] The initial biological sequence refers to unverified and unprocessed biological sequence information extracted from documents in the biomedical field. Sequence attribute information refers to relevant information in biomedical documents that is associated with the initial biological sequence and used to describe the sequence attributes of the initial biological sequence. For example, sequence attribute information may include a first sequence length and / or a first sequence type, where the first sequence length refers to the number of bases or amino acids in the biological sequence (such as any value among "5, 10, 15", subject to actual conditions, and this application does not specifically limit this), and the first sequence type refers to the molecular type of the biological sequence, which may include three types: DNA, RNA, or PRT (protein).

[0062] Understandably, to obtain the initial biological sequence and sequence attribute information from biomedical documents, when the document is an image, OCR technology can be used to perform text recognition and extract the text content from the image to obtain the initial biological sequence and its associated sequence attribute information. When the biomedical document includes both images and readable text, OCR technology can be used to extract the text content from the image and directly read the text content from the readable text to obtain the initial biological sequence and its associated sequence attribute information.

[0063] In this embodiment of the invention, different technologies are used to extract text content for different document formats in a targeted manner, which can ensure the accuracy of the extracted text content.

[0064] In one embodiment, the biomedical document can be an image. Specifically, the initial biological sequence and sequence attribute information can be obtained through the following steps, namely, step 110 specifically includes:

[0065] Step 111: Obtain documents related to the biomedical field;

[0066] Step 112: In the case where the document in the biomedical field includes images, perform text recognition on the biomedical field document to obtain multi-line recognized text;

[0067] Step 113: Classify the multi-line recognition text separately to obtain the text type of each line of recognition text, and determine the initial biological sequence and sequence attribute information from the multi-line recognition text based on the text type;

[0068] Step 114: Based on the sequence identifier corresponding to the initial biological sequence and the sequence identifier corresponding to the sequence attribute information, determine the sequence attribute information associated with the initial biological sequence.

[0069] Specifically, biomedical documents can be obtained from databases containing a large amount of biological sequence information, or from patent databases. For example, biomedical documents can be obtained from bioinformatics databases such as NCBI (National Center for Biotechnology Information), EMBL (European Molecular Biology Laboratory), and DDBJ (DNA Data Bank of Japan), or from patent databases such as Incopat and Patbase. This embodiment of the invention does not specifically limit the scope of these documents.

[0070] The aforementioned multi-line recognized text refers to the text content extracted from biomedical documents after text recognition using OCR technology. Sequence identifiers are unique identifiers used to identify biological sequences.

[0071] For example, Figure 2 This is a schematic diagram of a portion of the biological sequence in the sequence listing of a patent document in the biomedical field provided by this invention. The format of the sequence listing in the patent document conforms to the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listing Documents", such as... Figure 2 As shown, the elements enclosed in angle brackets "<>" are location identifiers, where the location identifier " <210> "" indicates a sequence identifier. In the sequence listing, each biological sequence has an independent and unique sequence identifier. The details of that biological sequence are located between one sequence identifier and the next. "Location identifier" <211> " represents the sequence length, <212> "Represents sequence type, <213> "Represents the organism, i.e., the name of the organism from which the sequence originates; in the location identifier" <400> Following this is the sequence identifier of the biological sequence, along with the location identifier. <210> The content that follows corresponds to the biological sequence, which begins on the next line, i.e., the nucleotide and / or amino acid sequence.

[0072] Figure 2The text content shown is a portion of the multi-line recognition text. After extracting the multi-line recognition text, a pre-trained deep learning-based text classification model can be used to classify the text, obtaining the text type of each line. Here, text classification can be performed based on keywords, for example, by using location identifiers to classify the multi-line recognition text, obtaining the text type of each line. This allows the determination of the initial biological sequence and sequence attribute information based on the text type. It can be understood that the first sequence length in the sequence attribute information is the location identifier. <211> "The corresponding sequence length, the first sequence type is the location identifier." <212> The corresponding sequence type.

[0073] Here, the aforementioned text type refers to a type of information used to characterize the specific content contained in each line of multi-line recognition text. For example, the text type can be a sequence identifier, sequence length, sequence type, etc. When the text type of the recognized text is a sequence identifier, sequence length, or sequence type, that is, the location identifier is " <210> "", <211> "or" <212> In the case of "", it can be determined that the identified text is sequence attribute information, and the text type of the identified text is a biological sequence, that is, the location identifier is "". <400> In the case of “”, it can be determined that the next line of text is the initial biological sequence.

[0074] Since each biological sequence has a unique sequence identifier in the sequence listing, after determining the initial biological sequence and sequence attribute information from multiple lines of recognized text, the initial biological sequence and sequence attribute information can be associated based on their respective sequence identifiers. When determining the initial biological sequence, the sequence identifier corresponding to each initial biological sequence, i.e., the location identifier, can be obtained. <400> The sequence number following the initial sequence number is used to match the sequence identifier with the sequence identifier in the sequence attribute information. If a match is successful, the sequence attribute information associated with the initial biological sequence can be determined. For example... Figure 2 As shown, the initial biological sequence A and sequence attribute information B can be determined from the multi-line recognition text. The sequence identifier corresponding to the initial biological sequence A is "1", and the sequence identifier corresponding to the sequence attribute information B is also "1". Therefore, based on this sequence identifier, the sequence attribute information B associated with the initial biological sequence A can be determined.

[0075] In another embodiment, biomedical documents may include images and readable text, for example, Figure 2The initial biological sequence A shown is in image format, and the sequence attribute information B is in readable text format. For the readable text, the sequence attribute information can be extracted according to the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listings". For the image, the text content can be extracted first using OCR technology, and then the initial biological sequence and its corresponding sequence identifier can be extracted from the text content according to the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listings". After obtaining the initial biological sequence, its corresponding sequence identifier, and the sequence attribute information, the initial biological sequence and the sequence attribute information can be associated based on the sequence identifier corresponding to the initial biological sequence and the sequence identifier corresponding to the sequence attribute information.

[0076] It should be noted that, in this embodiment of the invention, before identifying and extracting text content from an image using OCR technology, OCR can be trained using self-calibrated data. For example, it can be fine-tuned and optimized on patent data in the biomedical field, such as Chinese patent documents and European patent documents, to improve the accuracy of text recognition. The optimized OCR has a significant improvement in the accuracy of recognizing patent documents in the biomedical field.

[0077] In this embodiment of the invention, after the initial biological sequence and sequence attribute information are extracted, the initial biological sequence and the sequence attribute information are associated through the sequence identifier corresponding to the initial biological sequence and the sequence identifier corresponding to the sequence attribute information. This facilitates subsequent sequence attribute verification of the initial biological sequence through the correct sequence attribute information, thereby ensuring the accuracy and reliability of the attribute verification results.

[0078] Furthermore, obtaining documents in the biomedical field can be achieved through the following steps, specifically step 111, which includes:

[0079] Step 1111: Obtain the initial domain file;

[0080] Step 1112: If the initial domain file includes images, perform sequential page classification on the images to obtain the sequential page classification results of the images;

[0081] Step 1113: Based on the images that belong to the sequence page classification results and the readable text in the initial domain file, determine the biomedical domain file.

[0082] It should be noted that existing biological sequence identification and extraction technologies require document analysis and identification one by one, which is time-consuming and labor-intensive. In order to improve the efficiency of biological sequence identification and extraction, this embodiment of the invention filters the initial domain documents to obtain biomedical domain documents containing biological sequences, thereby improving the efficiency of biological sequence identification and extraction based on biomedical domain documents.

[0083] Here, the initial domain file refers to documents related to the biomedical field. It can include complete documents containing biological sequences within the biomedical field, while the biomedical domain file refers to the portion containing biological sequences obtained after filtering the initial domain file. For example, the initial domain file can be a patent document in the biomedical field, and the biomedical domain file can be the portion of that patent document containing biological sequences, i.e., the sequence page of that patent document. Compared to the initial domain file, the biomedical domain file contains less or even no redundant information; therefore, extracting biological sequences based on the biomedical domain file can improve efficiency.

[0084] Specifically, the initial domain name can be obtained from a database containing a large number of biomedical-related documents, or from a patent database. After obtaining the initial domain name, it can be determined whether the initial domain name includes images. If the initial domain name includes images, a sequence page filtering model can be used to classify the images into sequence pages to determine whether the images contain biological sequence information.

[0085] Furthermore, if the initial domain file includes images, it can be either an image PDF or a PDF containing readable text and images.

[0086] In one embodiment, the initial domain file can be an image PDF. The initial domain file is split page by page and input into a sequence page filtering model. The sequence page filtering model classifies each page of images into sequence pages, obtaining the image sequence page classification results. Here, sequence page classification refers to determining whether each page of images in the initial domain file is a sequence page using the sequence page filtering model, obtaining a sequence page classification result of "yes" or "no". If the sequence page classification result is "yes", it indicates that the image belongs to a sequence page, meaning that the image contains biological sequence information. After filtering out all images belonging to sequence pages in the initial domain file, the biomedical domain file can be determined based on these images. In this embodiment of the invention, by using a sequence page filtering model to filter out each page of images belonging to sequence pages from the initial domain file, thereby determining the biomedical domain file, the problem of redundant data consuming resources in the initial domain file can be solved.

[0087] In another embodiment, the initial domain file can be a PDF containing readable text. Since biological sequences may exist in the file as images, and the sequence characters cannot be directly read, a document parsing tool, such as a PDFparser, can be used to automatically extract the readable text and unreadable portions of the initial domain file. The unreadable portions are the images. In this case, a sequence page filtering model can be used to classify the images into sequence pages to determine whether an image belongs to a sequence page. After determining that an image belongs to a sequence page, the biomedical domain file can be identified based on the image and the extracted readable text. That is, the readable text present in the initial domain file can be directly used as the content of the biomedical domain file for subsequent biological sequence extraction.

[0088] Preferably, after filtering out the images belonging to sequence pages in the initial domain file using the sequence page filtering model, each image belonging to the sequence page can be optimized by denoising, resolution enhancement, etc., to obtain optimized sequence page images, thereby enabling more accurate initial biological sequences to be extracted based on these sequence page images.

[0089] It should be understood that the sequence page filtering model is a binary classification model used to determine whether an image in the initial domain file is a sequence page or a non-sequence page, thereby identifying the biomedical domain file based on the judgment result. Before performing step 1112, the sequence page filtering model can be pre-trained. Specifically, the sequence page filtering model can be trained as follows: collect a large number of sample documents in the biomedical domain. These sample documents can be scanned copies with image format pages. After splitting the sample documents page by page, manually label the positive and negative samples to obtain a positive sample set (containing multiple sequence pages) and a negative sample set (containing multiple non-sequence pages). Input the positive and negative sample sets into the initial binary classification model and iteratively optimize it to obtain the trained sequence page filtering model.

[0090] Step 120: Based on sequence attribute information, perform sequence attribute verification on the initial biological sequence to obtain the attribute verification results;

[0091] It should be noted that existing biological sequence identification technologies suffer from low accuracy. The extracted biological sequence information still requires manual intervention and calibration, which increases the workload. To address this, this invention automatically verifies the sequence attributes of the initial biological sequence based on the sequence attribute information associated with it, ensuring that the extracted biological sequence is more accurate, improving the accuracy of biological sequence identification and extraction, while reducing reliance on manual intervention and improving the efficiency of biological sequence extraction.

[0092] Specifically, sequence attribute verification refers to the process of verifying certain specific attributes of an initial biological sequence, where these specific attributes can be sequence length and / or sequence type. For example, the sequence attributes of the initial biological sequence can be predicted by analysis first, and then the sequence attributes included in the obtained sequence attribute information can be extracted to verify whether the predicted sequence attributes of the initial biological sequence are correct, thus ensuring that the accuracy of the finally extracted biological sequence is higher.

[0093] In one embodiment, the sequence attribute information may include a first sequence length and / or a first sequence type, and step 120 may be implemented through the following steps:

[0094] Step 121: Calculate the sequence length of the initial biological sequence as the second sequence length; perform sequence length verification on the initial biological sequence based on the second sequence length and the first sequence length to obtain the length verification result; and / or, perform sequence type verification on the initial biological sequence based on the first sequence type to obtain the type verification result.

[0095] Step 122: Determine the attribute validation result based on the length validation result and / or type validation result.

[0096] It should be noted that the second sequence length refers to the sequence length obtained based on the statistics of the initial biological sequence, while the first sequence length refers to the actual sequence length of the initial biological sequence obtained based on sequence attribute information. By comparing the difference between the second sequence length and the first sequence length, the sequence length can be verified and the length verification result can be obtained.

[0097] Specifically, when obtaining the second sequence length of the initial biological sequence, this can be achieved using a string length calculation function. The first sequence length can be directly obtained based on sequence attribute information, for example... Figure 2 By reading the location identifier " <211> The following content yields the first sequence lengths of the two biological sequences: "96" and "33". Comparing the second sequence length with the first sequence length, if the percentage difference between the second and first sequence lengths is greater than or equal to a preset percentage value, the length verification result is considered successful; if the percentage difference is less than the preset percentage value, the length verification result is considered unsuccessful.

[0098] Here, the preset percentage value can be set according to actual needs. For example, it can be set to 95%. If the length of the first sequence obtained based on sequence attribute information is 10000, and the length of the second sequence obtained based on the initial biological sequence is 9999, the percentage of the second sequence length to the first sequence length is 99.99%, exceeding the preset percentage value of 95%. Therefore, the length verification result can be determined as successful. Conversely, if the length of the first sequence obtained based on sequence attribute information is 10, and the length of the second sequence obtained based on the initial biological sequence is 9, the percentage of the second sequence length to the first sequence length is 90%, which is less than the preset percentage value of 95%. Therefore, the length verification result can be determined as failed.

[0099] It should be understood that the first sequence type refers to the true sequence type of the initial biological sequence obtained based on sequence attribute information. Different biological sequences have different characteristics. For example, DNA sequences are usually composed of four bases: A, T, C, and G; RNA sequences are usually composed of four bases: A, U, C, and G; and protein sequences are usually composed of 20 amino acids. Therefore, the initial biological sequence can be validated based on these characteristics to determine whether its composition conforms to the second sequence type in the sequence attribute information, thus obtaining the type validation result.

[0100] Preferably, in step 121, based on the first sequence type, the initial biological sequence is subjected to sequence type verification to obtain the type verification result, which can be achieved through the following steps:

[0101] Step 1211: Determine the first unit type of the basic units constituting the first sequence type of biological sequence, and determine the second unit type of each basic unit contained in the initial biological sequence;

[0102] Step 1212: Based on the first unit type and the second unit type, perform sequence type verification on the initial biological sequence to obtain the type verification result.

[0103] Specifically, a basic unit refers to the smallest unit that constitutes the initial biological sequence. In a protein sequence, the basic unit is an amino acid; in a DNA sequence, it is the four bases A, T, C, and G; and in an RNA sequence, it is the four bases A, U, C, and G. Based on the initial biological sequence, its constituent basic units can be extracted. Based on the first sequence type in the sequence attribute information, the basic unit of the biological sequence corresponding to that first sequence type can be determined. Then, it is determined whether the second unit type of each extracted basic unit matches the first unit type of the basic unit of the biological sequence corresponding to the sequence type. If they match, the sequence type verification is successful; otherwise, the sequence type verification fails.

[0104] For example, in a biomedical field document, based on the initial biological sequence, the extracted basic units only include various protein-related amino acids. Based on the sequence attribute information, the first sequence type obtained is "PRT", which means that the biological sequence corresponding to this sequence type is a protein sequence. Obviously, the second unit type of each extracted basic unit is consistent with the first unit type of the basic unit describing the biological sequence corresponding to the sequence type. Therefore, it can be determined that the sequence type verification result is successful.

[0105] As another example, in a biomedical field document, based on the initial biological sequence, the extracted basic units only include various protein-related amino acids, while the first sequence type obtained based on sequence attribute information is "DNA". Since the basic units corresponding to the DNA sequence should include four bases: A, T, C, and G, but the basic units extracted based on the initial biological sequence only include amino acids, it is obvious that the second unit type of the extracted basic units does not match the first unit type of the basic units of the biological sequence corresponding to the sequence type. Therefore, it can be determined that the sequence type verification result is a verification failure.

[0106] For another example, based on the initial biological sequence, the extracted basic units include four bases A, T, C, and G related to "DNA", and / or four bases A, U, C, and G related to "RNA", indicating that the second unit type is "DNA" and / or "RNA". However, if the first sequence type determined based on the sequence attribute information is "PRT", then it is obvious that the extracted second unit type does not match the recorded first unit type. Therefore, it can be determined that the sequence type verification result is a verification failure.

[0107] It should be noted that although the above example only illustrates how to compare the first unit type and the second unit type to determine the type verification result, those skilled in the art should understand that when determining the second unit type, there may be situations where the second unit type cannot be determined or is determined incorrectly due to character misidentification. In such cases, the attribute verification result is very likely to fail. Therefore, it is necessary to perform sequence correction on the initial biological sequence and then compare it again to avoid missing the correct sequence. The specific steps will be explained in detail below.

[0108] Furthermore, after obtaining the length verification result and / or type verification result, the attribute verification result can be determined. In one embodiment, the attribute verification result can be determined based on the length verification result, in which case the length verification result can be used as the attribute verification result. In another embodiment, the attribute verification result can be determined based on the type verification result, in which case the type verification result can be used as the attribute verification result. In yet another embodiment, the attribute verification result can be determined based on both the length verification result and the type verification result. If both the length verification result and the type verification result are successful, the attribute verification result can be determined as successful. If either the length verification result or the type verification result is unsuccessful, the attribute verification result can be determined as unsuccessful. It is understood that, in the embodiments of the present invention, according to actual application requirements, i.e., high requirements for sequence extraction accuracy, the attribute verification result can be determined simultaneously based on the length verification result and the type verification result, thereby determining the final extracted biological sequence based on the attribute verification result, which can effectively ensure the accuracy of the extracted biological sequence.

[0109] Furthermore, when determining the attribute validation result based on the length validation result and the type validation result, step 122 above specifically includes:

[0110] Step 1221: If the length verification result is successful and the type verification result is unsuccessful, perform sequence correction on the initial biological sequence based on the correction method corresponding to the first sequence type, and perform sequence attribute verification based on the sequence-corrected initial biological sequence to obtain the corrected length verification result and type verification result.

[0111] Step 1222: If both the corrected length verification result and the type verification result are successful, then the attribute verification result is determined to be successful.

[0112] It should be noted that, considering that the initial biological sequence obtained by text recognition may have problems such as partial content loss or format disorder, this embodiment of the invention considers the possibility of text recognition errors when the length verification result is successful but the type verification result is unsuccessful. Therefore, the initial biological sequence is corrected and the sequence attribute is verified again based on the corrected initial biological sequence, thereby ensuring the reliability and accuracy of the obtained attribute verification results and improving the accuracy of biological sequence recognition and extraction.

[0113] Specifically, sequence correction of the initial biological sequence can be based on the correction method corresponding to the first sequence type, which may include the following aspects: 1) Removing non-letter and non-space characters from the initial biological sequence; 2) For protein, DNA, and RNA sequences, summarizing common sequence recognition errors and replacing or correcting sequence recognition errors present in the initial biological sequence; 3) For protein sequences, regardless of case, replacing the amino acid representations in the initial biological sequence with standard representations, such as capitalizing the first letter; 4) For protein sequences, using a regular expression pattern that includes insertions / deletions / replacements and whose edit distance meets preset conditions (e.g., edit distance less than 1), segmenting the initial biological sequence, replacing recognition errors whose edit distance meets preset conditions, and deleting non-matching single letters.

[0114] Based on the sequence correction method described above, prior knowledge can be summarized, for example, a pre-defined standard sequence unit dictionary can be constructed. If the length verification result is successful but the type verification result is unsuccessful, the initial biological sequence is traversed, and characters with incorrect recognition are corrected according to the pre-defined standard sequence unit dictionary, resulting in a corrected initial biological sequence. For the corrected initial biological sequence, the sequence attribute verification step can be repeated to obtain corrected length and type verification results. Based on the corrected length and type verification results, the accurate attribute verification result can be determined, and thus, based on the attribute verification result, it can be determined whether the initial biological sequence should be retained and used as the target biological sequence.

[0115] For example, glutamic acid (Glu) contains a lowercase letter "l", which is often recognized as the Arabic numeral "1" during character recognition. Therefore, when performing sequence correction on an initial biological sequence containing glutamic acid (Glu), the incorrectly recognized character can be automatically corrected based on prior knowledge. Similarly, threonine (Thr) might be incorrectly recognized as "Th" during character recognition using OCR technology. Since threonine (Thr) is the only amino acid starting with "Th" in the standard sequence unit dictionary, it can be automatically corrected.

[0116] Understandably, after performing sequence attribute verification on the corrected initial biological sequence again to obtain the corrected length verification result and type verification result, if both the length verification result and the type verification result are successful, then the attribute verification result can be determined to be successful; if either the length verification result or the type verification result is unsuccessful, then the attribute verification result can be determined to be unsuccessful.

[0117] Step 130: Extract the initial biological sequence whose attribute verification result is successful, and use it as the target biological sequence.

[0118] Specifically, based on sequence attribute information, after verifying the sequence attributes of the initial biological sequence, an attribute verification result can be obtained. If the attribute verification result is successful, it indicates that the extracted initial biological sequence is accurate and complete, and therefore, it can be used as the target biological sequence. If the attribute verification result is unsuccessful, it indicates that the extracted initial biological sequence has missing, inaccurate, or incomplete parts, and therefore, it can be discarded, thus ensuring that the final extracted target biological sequence has higher accuracy and reliability.

[0119] The method provided in this invention obtains an initial biological sequence from a biomedical document and sequence attribute information associated with the initial biological sequence. Based on the sequence attribute information, the method verifies the sequence attributes of the initial biological sequence to extract the successfully verified initial biological sequence as the final target biological sequence. This ensures a high accuracy rate for biological sequence identification and extraction, and eliminates the need for manual intervention and calibration, further improving the efficiency of biological sequence extraction. Compared with existing technologies, this invention determines the target biological sequence based on attribute verification results, resulting in higher accuracy and reliability of the extracted biological sequence.

[0120] Based on any of the above embodiments Figure 3 This is the second flowchart of the biological sequence extraction method provided by the present invention, as shown below. Figure 3 As shown, the method includes:

[0121] S1, Obtain the initial domain file;

[0122] Specifically, the initial field document can be a patent PDF file containing biological sequences in the biomedical field, which can be obtained from a patent database. There can be one or more initial field documents, and this embodiment of the invention does not specifically limit this.

[0123] S2, when the initial domain file is an image PDF, the initial biological sequence and the sequence attribute information associated with the initial biological sequence can be obtained in the following way:

[0124] S2-1: Based on the trained sequence page filtering model, classify the sequence pages of the image PDF, extract each sequence page in the image PDF, optimize each sequence page, and determine the biomedical field documents based on the optimized sequence pages.

[0125] S2-2 uses OCR technology to extract the initial biological sequence from documents in the biomedical field, as well as the sequence attribute information associated with the initial biological sequence;

[0126] Specifically, the sequential page selection model is a binary classification model used to determine whether image pages in an image PDF are sequential pages or non-sequential pages. The sequential page selection model can be trained as follows: Collect a large number of sample documents in the biomedical field. These sample documents can be scanned copies with image format pages. After splitting the sample documents page by page, manually label the positive and negative samples to obtain a positive sample set (containing multiple sequential pages) and a negative sample set (containing multiple non-sequential pages). Input the positive and negative sample sets into the initial binary classification model and iteratively optimize it to obtain the trained sequential page selection model.

[0127] After filtering out each sequence page in an image PDF using a trained sequence page filtering model, each sequence page can be optimized by denoising, resolution enhancement, and other processes to obtain optimized sequence pages, which can then be used as files in the biomedical field.

[0128] OCR technology was used to perform text recognition on documents in the biomedical field, resulting in multi-line recognized text. Following the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listing Documents," the initial biological sequence, its corresponding sequence identifier, and sequence attribute information were extracted from the multi-line recognized text. The sequence attribute information may include the sequence identifier, sequence length, and sequence type. Based on the sequence identifier corresponding to the initial biological sequence and the sequence identifiers corresponding to the sequence attribute information, the sequence attribute information associated with the initial biological sequence can be determined.

[0129] S3, assuming the initial domain file is a readable PDF, the initial biological sequence and associated sequence attribute information can be obtained in the following ways:

[0130] S3-1: Using the PDF parser tool, readable text and sequence images are extracted from the initial domain files. Based on the readable text and sequence images, the biomedical domain files are identified.

[0131] S3-2, based on readable text, sequence attribute information is read, and the initial biological sequence in the sequence image is extracted using OCR technology;

[0132] Specifically, the initial domain file is a patent PDF file. If the initial domain file is not an image PDF, it can be determined that it is a readable PDF. Here, a readable PDF refers to a patent PDF file containing readable text. In the sequence listing section of the patent PDF file, the sequence attribute information can be readable text, and its text content can be directly read. The initial biological sequence is an image, and its content cannot be directly read. For readable text, the sequence attribute information can be read according to the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listings." This sequence attribute information can include sequence identifiers, sequence lengths, and sequence types. For sequence images, the text content can first be extracted from the sequence image using OCR technology, and then the initial biological sequence and its corresponding sequence identifier can be extracted from the text content according to the format listed in the "Standard for Nucleotide and / or Amino Acid Sequence Listings and Electronic Sequence Listings." Based on the sequence identifiers corresponding to the initial biological sequence and the sequence identifiers corresponding to the sequence attribute information, the sequence attribute information associated with the initial biological sequence can be determined.

[0133] S4. Based on the sequence attribute information associated with the initial biological sequence, the sequence attribute of the initial biological sequence is verified. The sequence attribute information may include a sequence identifier, sequence length, and sequence type. The sequence identifier can be associated with the correct biological sequence and stored in the database to facilitate later locating the source of the biological sequence. The sequence length and sequence type can be used to verify the sequence attributes of the initial biological sequence. The specific verification method is as follows:

[0134] S4-1, Analyze the extracted initial biological sequence, calculate its sequence length, and obtain the second sequence length, which can be denoted as A. Take the sequence length in the sequence attribute information as the first sequence length, which can be denoted as B. Compare the difference between A and B. If the preset percentage value is met, the initial biological sequence is retained and proceeds to step S4-2 for further processing; if the preset percentage value is not met, the initial biological sequence is discarded.

[0135] S4-2. Because different biological sequences have different characteristics—for example, DNA sequences are usually composed of four bases (A, T, C, G), RNA sequences are usually composed of four bases (A, U, C, G), and protein sequences are usually composed of 20 amino acids—the initial biological sequence can be validated based on these characteristics to determine whether its composition conforms to the first sequence type in the sequence attribute information. If it does, the initial biological sequence can be retained, the attribute validation result is confirmed as successful, and the process proceeds to step S5. If it does not conform, the possibility of an OCR recognition error is considered, and the process proceeds to step S4-3 for further processing.

[0136] S4-3: Based on prior knowledge, such as a preset standard sequence unit dictionary, traverse the initial biological sequence to correct misidentified characters in the initial biological sequence and obtain a corrected initial biological sequence. The corrected initial biological sequence is then processed again in step S4 to obtain the corrected length verification result and type verification result. If both the corrected length verification result and type verification result are successful, the attribute verification result is considered successful, and the corresponding initial biological sequence can be retained. If either the corrected length verification result or type verification result is considered unsuccessful, the attribute verification result is considered unsuccessful, and the corresponding initial biological sequence can be discarded.

[0137] S5, extract the initial biological sequence that has been successfully verified as the attribute verification result, and use it as the target biological sequence.

[0138] Based on any of the above embodiments Figure 4 This is a flowchart illustrating the biological sequence retrieval method provided by the present invention, as shown below. Figure 4 As shown, taking the application of this method to a terminal as an example, the method includes:

[0139] Step 410: Obtain the target search terms;

[0140] Step 420: Using a biological sequence library, determine the target biological sequence that matches the target search term. The biological sequence library is constructed based on the target biological sequence determined by any of the biological sequence extraction methods described above.

[0141] Step 430: Display the target biological sequence that matches the target search term.

[0142] Specifically, following the biological sequence extraction method described in the above embodiments, after obtaining the biological sequences contained in biomedical documents, a highly reliable biological sequence library can be constructed to facilitate users' rapid retrieval of target biological sequences. For example, after extracting the biological sequences, the text format of the biological sequences can be saved as the FASTA sequence standard format, and a database that can be used for biological sequence retrieval, i.e., a biological sequence library, can be constructed based on this. It is understood that when constructing the biological sequence library, the sequence identifier in the sequence attribute information can be associated with the correct biological sequence and stored in the biological sequence library to facilitate rapid location of the source of the target biological sequence during retrieval.

[0143] Users can input target search terms via smartphones, computers, tablets, or other terminals and send the target search terms to a server for retrieval. Here, the target search term can be the complete form of the target biological sequence or a truncated form of the target biological sequence; this embodiment of the invention does not specifically limit this. Users can input the target search term directly or upload a file containing the target search term; this embodiment of the invention does not specifically limit this either.

[0144] Upon receiving the target search term, the system can identify the corresponding biological sequence in the biological sequence database and push this sequence to the terminal for display and user viewing. It should be noted that after receiving the target search term, the BLAST (Basic Local Alignment Search Tool) algorithm can be used to perform a search and obtain the corresponding biological sequence. Furthermore, the terminal and server mentioned above can communicate via a network, including but not limited to wide area networks (WANs), metropolitan area networks (MANs), or local area networks (LANs).

[0145] The biological sequence retrieval method provided in this invention constructs a highly reliable biological sequence library based on biological sequences extracted from documents in the biomedical field. This library enables rapid retrieval and querying of biological sequences. Before screening or designing a biopharmaceutical, companies can quickly search for the biological sequences of known similar biopharmaceuticals based on this library, thereby reducing the risk of infringement, facilitating timely adjustments to R&D strategies, and saving R&D costs.

[0146] Based on any of the above embodiments Figure 5 This is a schematic diagram of the biological sequence extraction device provided by the present invention, as shown below. Figure 5 As shown, the device includes: an acquisition unit 510, used to acquire an initial biological sequence from a document in the biomedical field, and sequence attribute information associated with the initial biological sequence; a verification unit 520, used to perform sequence attribute verification on the initial biological sequence based on the sequence attribute information, and obtain attribute verification results; and an extraction unit 530, used to extract the initial biological sequence whose attribute verification result is successful, as the target biological sequence.

[0147] The biological sequence extraction device provided in this invention obtains an initial biological sequence from a biomedical document and sequence attribute information associated with the initial biological sequence. Based on this sequence attribute information, the device verifies the sequence attributes of the initial biological sequence to extract the successfully verified initial biological sequence as the final target biological sequence. This ensures high accuracy in biological sequence identification and extraction, and eliminates the need for manual intervention and calibration, further improving the efficiency of biological sequence extraction. Compared with existing biological sequence identification and extraction technologies, this invention determines the target biological sequence based on attribute verification results, resulting in higher accuracy and reliability of the extracted biological sequence.

[0148] Based on any of the above embodiments, the sequence attribute information includes a first sequence length and / or a first sequence type. The verification unit 520 specifically includes: an attribute verification subunit, used to count the sequence length of the initial biological sequence as a second sequence length, and to perform sequence length verification on the initial biological sequence based on the second sequence length and the first sequence length to obtain a length verification result; and / or, to perform sequence type verification on the initial biological sequence based on the first sequence type to obtain a type verification result; and a result determination subunit, used to determine the attribute verification result based on the length verification result and / or the type verification result.

[0149] Based on any of the above embodiments, the attribute verification subunit is specifically used to: determine the first unit type of the basic units constituting the first sequence type of the biological sequence, and determine the second unit type of each basic unit contained in the initial biological sequence; and perform sequence type verification on the initial biological sequence based on the first unit type and the second unit type to obtain the type verification result.

[0150] Based on any of the above embodiments, the result determination subunit is specifically used for: when the length verification result is successful and the type verification result is unsuccessful, performing sequence correction on the initial biological sequence based on the correction method corresponding to the first sequence type, and performing sequence attribute verification based on the sequence-corrected initial biological sequence to obtain the corrected length verification result and type verification result; when both the corrected length verification result and type verification result are successful, determining the attribute verification result as successful.

[0151] Based on any of the above embodiments, the acquisition unit 510 specifically includes: a file acquisition subunit, used to acquire biomedical field documents; a text recognition subunit, used to perform text recognition on biomedical field documents when the biomedical field documents include images, to obtain multi-line recognized text; a text classification subunit, used to classify the multi-line recognized text separately, to obtain the text type of each line of recognized text, and to determine the initial biological sequence and sequence attribute information from the multi-line recognized text based on the text type; and an association subunit, used to determine the sequence attribute information associated with the initial biological sequence based on the sequence identifier corresponding to the initial biological sequence and the sequence identifier corresponding to the sequence attribute information.

[0152] Based on any of the above embodiments, the file acquisition subunit is specifically used for: acquiring an initial domain file; if the initial domain file includes images, performing sequence page classification on the images to obtain the sequence page classification result of the images; and determining the biomedical domain file based on the sequence page classification result of the images belonging to the sequence page and the readable text in the initial domain file.

[0153] Based on any of the above embodiments Figure 6 This is a schematic diagram of the biological sequence retrieval device provided by the present invention, as shown below. Figure 6 As shown, the device includes: a search term acquisition unit 610 for acquiring target search terms; a sequence retrieval unit 620 for determining target biological sequences that match the target search terms through a biological sequence library, wherein the biological sequence library is constructed based on target biological sequences determined by any of the biological sequence extraction methods described above; and a sequence display unit 630 for displaying target biological sequences that match the target search terms.

[0154] The biological sequence retrieval device provided in this invention constructs a highly reliable biological sequence library based on biological sequences extracted from documents in the biomedical field. It can provide rapid retrieval and query of biological sequences. Before screening or designing a biopharmaceutical, companies can quickly search for the biological sequences of known similar biopharmaceuticals based on this biological sequence library, thereby reducing the company's infringement risk, facilitating timely adjustment of R&D strategies, and saving R&D costs.

[0155] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7As shown, the electronic device may include a processor 710, a communications interface 720, a memory 730, and a communication bus 740, wherein the processor 710, communications interface 720, and memory 730 communicate with each other via the communication bus 740. The processor 710 can invoke logical instructions in the memory 730 to execute a biological sequence extraction method, which includes: acquiring an initial biological sequence from a biomedical document, and sequence attribute information associated with the initial biological sequence; performing sequence attribute verification on the initial biological sequence based on the sequence attribute information, and obtaining an attribute verification result; extracting the initial biological sequence whose attribute verification result is successful as the target biological sequence. The processor 710 can also invoke logical instructions in the memory 730 to execute a biological sequence retrieval method, which includes: acquiring a target search term; determining a target biological sequence matching the target search term using a biological sequence library, the biological sequence library being constructed based on the target biological sequences determined by any of the above biological sequence extraction methods; and displaying the target biological sequence matching the target search term.

[0156] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0157] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the biological sequence extraction method provided by the above methods. The method includes: acquiring an initial biological sequence from a document in the biomedical field, and sequence attribute information associated with the initial biological sequence; performing sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result; and extracting the initial biological sequence whose attribute verification result is successful as the target biological sequence.

[0158] Furthermore, when the computer program is executed by the processor, the computer is able to perform the biological sequence retrieval methods provided by the above methods, which include: obtaining target search terms; determining target biological sequences that match the target search terms through a biological sequence library, wherein the biological sequence library is constructed based on the target biological sequences determined by any of the above biological sequence extraction methods; and displaying the target biological sequences that match the target search terms.

[0159] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program is implemented to perform the biological sequence extraction method provided by the above methods. The method includes: acquiring an initial biological sequence from a document in the biomedical field, and sequence attribute information associated with the initial biological sequence; performing sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result; and extracting the initial biological sequence whose attribute verification result is successful as the target biological sequence.

[0160] When executed by a processor, the computer program implements the biological sequence retrieval methods provided by the methods described above, the methods including: obtaining target search terms; determining target biological sequences that match the target search terms through a biological sequence library, the biological sequence library being constructed based on the target biological sequences determined by any of the biological sequence extraction methods described above; and displaying the target biological sequences that match the target search terms.

[0161] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0162] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of biological sequence extraction, characterized by, The method comprises the following steps: obtaining an initial biological sequence in a biological medicine field file and sequence attribute information associated with the initial biological sequence; performing sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result; extracting the initial biological sequence with a successful attribute verification result as the attribute verification result as a target biological sequence; wherein the sequence attribute information comprises a first sequence length and / or a first sequence type, the first sequence length being the actual sequence length of the initial biological sequence, and the first sequence type being the actual sequence type of the initial biological sequence; the sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result comprises: statistically obtaining a sequence length of the initial biological sequence as a second sequence length, performing sequence length verification on the initial biological sequence based on the second sequence length and the first sequence length to obtain a length verification result, and / or performing sequence type verification on the initial biological sequence based on the first sequence type to obtain a type verification result; determining the attribute verification result based on the length verification result and / or the type verification result; the determination of the attribute verification result based on the length verification result and / or the type verification result comprises: in the case that the length verification result is successful and the type verification result fails, determining that the initial biological sequence has character recognition errors, correcting the character recognition errors in the initial biological sequence based on a correction method corresponding to the first sequence type, and performing sequence attribute verification on the corrected initial biological sequence again, and if the verification is successful, determining that the attribute verification result is successful.

2. The biological sequence extraction method of claim 1, wherein, the sequence type verification on the initial biological sequence based on the first sequence type to obtain a type verification result comprises: determining a first unit type of a basic unit of a biological sequence constituting the first sequence type, and determining a second unit type of each basic unit contained in the initial biological sequence; performing sequence type verification on the initial biological sequence based on the first unit type and the second unit type to obtain the type verification result.

3. The biological sequence extraction method according to claim 1 or 2, characterized in that, The method comprises the following steps: obtaining a biological medicine field file; in the case that the biological medicine field file comprises an image, performing character recognition on the biological medicine field file to obtain a plurality of lines of recognized text; performing text classification on the plurality of lines of recognized text to obtain a text type of each line of recognized text, and determining the initial biological sequence and the sequence attribute information from the plurality of lines of recognized text based on the text type; determining the sequence attribute information associated with the initial biological sequence based on a sequence identifier corresponding to the initial biological sequence and a sequence identifier corresponding to the sequence attribute information.

4. The biological sequence extraction method of claim 3, wherein, The method comprises the following steps: obtaining an initial field file; In a case that the initial field file comprises an image, performing sequence page classification on the image to obtain a sequence page classification result of the image; Based on the sequence page classification result that the image belongs to a sequence page, and readable text in the initial field file, determining the biomedical field file.

5. A method of biological sequence search, characterized by, Comprise: Acquiring a target search term; Determining a target biological sequence matched with the target search term through a biological sequence library, the biological sequence library being constructed based on the target biological sequence determined by the biological sequence extraction method in any one of claims 1 to 4; Displaying the target biological sequence matched with the target search term.

6. A biological sequence extraction apparatus characterized by comprising: Comprise: An acquisition unit configured to acquire an initial biological sequence in a biomedical field file and sequence attribute information associated with the initial biological sequence; A verification unit configured to perform sequence attribute verification on the initial biological sequence based on the sequence attribute information to obtain an attribute verification result; An extraction unit configured to extract the initial biological sequence with a verification success in the attribute verification result as a target biological sequence; The sequence attribute information comprises a first sequence length and / or a first sequence type, the first sequence length being a real sequence length of the initial biological sequence, and the first sequence type being a real sequence type of the initial biological sequence; The verification unit is specifically configured to: Statistically acquire a sequence length of the initial biological sequence as a second sequence length, perform sequence length verification on the initial biological sequence based on the second sequence length and the first sequence length to obtain a length verification result, and / or perform sequence type verification on the initial biological sequence based on the first sequence type to obtain a type verification result; Determine the attribute verification result based on the length verification result and / or the type verification result; The verification unit is further configured to: In a case that the length verification result is a verification success and the type verification result is a verification failure, determine that the initial biological sequence has character recognition errors, correct the character recognition errors in the initial biological sequence based on a correction mode corresponding to the first sequence type, and perform sequence attribute verification on the corrected initial biological sequence again, and if the verification is successful, determine that the attribute verification result is a verification success.

7. A biological sequence search device characterized by comprising: Comprise: A search term acquisition unit configured to acquire a target search term; A sequence retrieval unit configured to determine a target biological sequence matched with the target search term through a biological sequence library, the biological sequence library being constructed based on the target biological sequence determined by the biological sequence extraction method in any one of claims 1 to 4; A sequence display unit configured to display the target biological sequence matched with the target search term.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the biological sequence extraction method in any one of claims 1 to 4 or the biological sequence retrieval method in claim 5 when executing the program.

Images