55 results about "Biosequence" patented technology

Provided herein are technologies for framing and evaluating biological sequence-based analysis tasks in a unified, natural-language-based, text in and text out format. Among other things, methods and systems of the present disclosure provide machine-learning technologies for combining biological sequence data, representing, for example, DNA, RNA, and protein sequences, with natural language, conversational style prompts that set out particular analysis tasks to be performed on the biological sequence data. This approach, for example, allows complex analysis tasks, including, but not limited to, identification of various sequence modifications, genes, and regulatory elements in DNA sequences, and quantification of properties such as degradation propensity of RNA and protein stability, to be input to a machine learning model in a uniform text-based format and for output to be generated in a same, unified, text-based format.

The invention provides a heuristic biological sequence clustering method based on semi-global comparison. The method comprises the following steps: firstly, reading all biological sequences, removing repetitive sequences, and carrying out descending sorting on the biological sequences according to sequence lengths; a first sequence is used as a representative sequence of a first category, then a next sequence is read, the similarity between the current sequence and the representative sequence is calculated by adopting a semi-global sequence comparison algorithm, each sequence can be compared to the most similar area in the representative sequence through semi-global sequence comparison, the similarity between the sequences is found to the maximum extent, and the similarity between the sequences and the representative sequence is calculated. Obtaining a relatively high comparison similarity value between the sequences; if the similarity meets a clustering threshold value, adding the similarity into a category with the same representative sequence, otherwise, taking the similarity as a new representative sequence, and generating a new category; repeating the steps until all the sequences are processed; the number of the final representative sequences is the number of clustering categories, and the categories which are the same as the representative sequences are member sequences of each category.

The invention relates to the technical field of health evaluation, in particular to a group image health evaluation method and system based on intestinal flora, and the method comprises the steps: collecting intestinal flora samples of a target group through a standardized process, and obtaining microbial sequence information through a high-throughput sequencing technology; microflora diversity characteristics and core flora characteristics are extracted through strain identification and relative abundance analysis, individual flora characteristics are compared with a healthy population reference database, health state scores are calculated, health grades are divided, population health grade distribution is counted, and the population health grade distribution is calculated. And generating a group health evaluation report containing a visual chart and text analysis. According to the method, a complete technical system from sample collection to health strategy making is established, systematic evaluation of the group health state from the perspective of intestinal flora is achieved, and the limitation of a traditional method in group health early evaluation is overcome. The method is suitable for health monitoring of different scales of groups, and provides effective technical support for public health management and health service.

The invention belongs to the technical field of microbiome clinical diagnosis, and particularly relates to a marker combination, a model and a method for predicting the prognosis risk of a colorectal cancer patient. The method comprises the steps of whole genome sequencing data analysis, human source sequence filtering, microorganism sequence extraction and annotation and machine learning diagnosis model construction. The method has the advantages of being high in accuracy and free of culture, tissue-specific microbiota still exist in low-biomass tumors, and promising diagnostic value is shown. A large amount of clinical patient sequencing data can be subjected to secondary development and utilization, the accuracy of tumor diagnosis is improved, and the method can be widely applied to prognosis survival evaluation of patients.

The invention provides a dynamic multi-modal biological sequence staged pre-training system for precise breeding, which comprises a multi-modal data organization unit used for collecting a single-modal sequence, generating a two-modal pairing sequence, constructing a three-modal interpenetrating sequence and simulating biological information transmission; the unified sequence representation unit is used for performing unified word segmentation on each modal sequence and adding modal marks; the progressive pre-training strategy unit is used for training in three stages and dynamically adjusting the modal mixing proportion in combination with a simulated annealing strategy; the cross-modal autoregression unit realizes prediction conversion between modals; the sequence feature extraction and prediction unit is used for obtaining a quantitative prediction value. The invention further provides a dynamic multi-mode biological sequence staged pre-training method for precise breeding. Therefore, the biological sequence feature prediction precision can be remarkably improved, the training convergence speed is increased, the performance equivalent to that of a large model is achieved, the cost is low, flexible deployment in various environments can be achieved, and efficient, accurate and low-cost technical support is provided for biological sequence analysis.

Described are systems and methods for analyzing ultra-long context biological sequence data, comprising processing the biological sequence data using a machine learning model.

Compressing files is disclosed. An DNA sequence to be compressed is first aligned. Aligning the DNA sequence includes splitting the DNA sequences into smaller sequences or portions that can be aligned. After the DNA sequence is spilt one or more time and aligned, a compression matrix is generated. Each row of the compression matrix corresponds to part of the DNA sequence. A consensus sequence is determined from the compression matrix. Using the consensus sequence, pointer pairs are generated. Each pointer pair identifies a subsequence of the consensus matrix. The compressed file includes the pointer pairs and the consensus sequence.

This invention discloses a method for constructing an online prediction website based on cis-regulatory elements of biological sequences. This method is applicable to building online prediction websites from models trained using machine learning on biological sequences. With the continuous development of artificial intelligence technology, machine learning algorithms have flourished and been successfully applied in multiple fields. In particular, in the fields of biological and medical research, researchers have constructed numerous prediction models to serve various research tasks, such as the prediction of cis-regulatory elements, biomedical pathogenic elements, and proteins. Although many prediction models have been successfully constructed, professionals who frequently use computers are more concerned with how to apply these models to practical work. Therefore, constructing the prediction models in this paper into user-friendly and easy-to-use online websites can greatly enhance the practical application value of the models.

A method for facilitating rapid genome sequence analysis includes accessing an output stream of an alignment process that includes aligned reads of a biological sequence that are aligned to a reference genome. The method also includes distributing the aligned reads to a plurality of computing nodes based on genomic position. Each of the plurality of computing nodes is assigned to a separate data bin of a plurality of data bins associated with genomic position. The method also includes, for at least one aligned read determined to overlap separate data bins of the plurality of data bins, duplicating the at least one aligned read and distributing the at least one aligned read to separate computing nodes of the plurality of computing nodes that are assigned to the separate data bins.

The invention belongs to the technical field of bioinformatics, and particularly relates to an lncRNA and disease association prediction method based on deep learning, which comprises the following specific steps: firstly, according to known lncRNA-disease association information, disease-miRNA association information and lncRNA-miRNA association information, constructing an lncRNA-disease association matrix, a disease-miRNA association matrix and an lncRNA-miRNA association matrix; and constructing an lncRNA comprehensive similarity matrix LS, a disease comprehensive similarity matrix DS and a miRNA comprehensive similarity matrix MS. The biological sequence selective compression network adopted by the invention can effectively model a long-distance dependency relationship in an ultra-long sequence and dynamically distinguish key functional fragments from redundant fragments according to context.

The present invention provides a method for training a biological model and a method for processing a biological sequence, which can be applied to the field of artificial intelligence technology. The training method of the biological model includes: constructing a protein-ribonucleic acid sequence pair according to a preset input format based on a protein sequence data set and a ribonucleic acid sequence data set, and the protein-ribonucleic acid sequence pair includes a preset prompt word; using a preset word segmenter to perform word segmentation processing on the protein sequence data set and the ribonucleic acid sequence pair to obtain a training sequence pair; inputting the training sequence pair into the biological model to obtain a predicted sequence pair corresponding to the training sequence pair; training the biological model according to the predicted sequence pair, the training sequence pair and the preset prompt word to obtain a trained biological model. This method can realize the simultaneous understanding and generation of protein and ribonucleic acid sequences, providing a basis for downstream application docking.

The invention discloses a synthetic biological data intelligent extraction system based on a multi-modal large language model, and belongs to the technical field of biological data intelligent extraction, and the system comprises a multi-modal data preprocessing and standardization module which is used for carrying out the analysis, classification and structured output of input heterogeneous synthetic biological data; and the multi-modal large language model processing module is connected with the preprocessing module and is used for receiving the processed structured data and carrying out joint understanding and information extraction through knowledge enhancement and deep fusion technologies. The heterogeneous synthetic biological data can be efficiently processed, different types of data such as scientific literatures, chart images and biological sequences are accurately analyzed and classified through the multi-modal data preprocessing and standardization module, the data format barrier is broken, scattered biological information is systematically integrated, and the data processing efficiency is improved. A standard and unified data foundation is laid for subsequent information extraction, and the comprehensiveness and effectiveness of data processing are greatly improved.

Disclosed are a method for constructing a database, a method for labeling an association degree of biological sequences, a method for retrieving a document, and a computer device. In the solution of this application, a biological sequence and attribute information are extracted from a target document, and an entry in a database is constructed based on the extracted biological sequence and the attribute information. When a user conducts retrieval based on the database, a server can match an entry for the user by means of the biological sequence and the attribute information in the entry or a combination of the two. Therefore, when applied to a retrieval platform, the database of this application can provide the user with various types of retrieval support, such as biological sequence retrieval, biological sequence attribute retrieval, and comprehensive biological sequence and biological sequence attribute retrieval, and the like.

The application provides a roughness evaluation method and device of fitness space, electronic equipment and storage medium, and relates to the technical field of biology. The method comprises the following steps: acquiring a data set; the data set comprises a plurality of biological sequences; based on the mutation sites between each biological sequence in the data set, a plurality of adjacent sequence pairs in the data set are determined; the mutation sites between the two biological sequences in each adjacent sequence pair meet the set number of mutation sites; and the roughness of the fitness space is estimated according to each adjacent sequence pair, to obtain a roughness evaluation result. The application solves the problem of inaccurate roughness evaluation of the fitness space in the related art.

The present invention provides a method and apparatus for biosequence engineering guided by machine learning. Specifically, a Bayesian optimization-guided evolutionary algorithm (BO-EVO) is provided, which combines Bayesian optimization (BO) and evolutionary algorithm (EVO). EVO solves the problem of excessive computational complexity caused by BO locating the global optimum by brute force search of the entire design space. At the same time, the exploratory nature of BO is used to neutralize the greed and under-exploration shortcomings of EVO to achieve efficient iteration between machine learning models and robotic experiments, so as to economically obtain new protein variants of high value. The method of the present invention is expected to achieve efficient and scalable computing and exploration, and provide an efficient biosequence engineering solution.

This invention discloses a phylogenetic tree construction method and system based on deep learning and beam search. It predefines evolutionary scenarios, setting evolutionary parameters for each scenario with reference to real-world biological sequence attributes. Training, validation, and test sets are created based on simulated phylogenetic trees and corresponding multiple sequence alignment data according to the predefined parameters. A deep learning classifier with convolutional neural networks and long short-term memory neural networks as its core is constructed. The deep learning classifier is trained and validated using the training and validation sets, and its accuracy is tested using the test set. Based on the trained deep learning classifier and the sliding window method, classification predictions are performed on all sub-quad-sequence trees of the four-sequence data. A phylogenetic tree reconstruction is performed on the multiple sequence data using an improved stepwise addition method and the quad-sequence tree classification prediction results, resulting in a complete reconstruction. This enables phylogenetic tree construction under conditions of different species numbers and sequence lengths.

The present invention discloses a compression method and device for biological sequence identifiers, as well as a decompression method and device. For each identifier in a gene sequencing file, the identifier is split into several sub-identifiers; encoding rules for several windows are defined, and the encoding rules match the text format of the sub-identifiers; the sub-identifiers with the same referential meaning are divided into the same window; for each window, all the sub-identifiers in the window are encoded according to the corresponding encoding rule, and the encoding results of each window are aggregated into the compression result of the identifier. These methods maximize the compression ratio of all identifier data while ensuring compatibility with special data as much as possible, and at the same time ensure the encoding and decoding performance.

The invention discloses a method for obtaining and correcting biological sequence information, which comprises the following steps of: performing two rounds of different bit sequencing on a sample to be detected, introducing supervisory information, and realizing accurate decoding on the sequence information to be detected. The defect that signal decoding of two rounds of bit sequencing is sensitive to noise and the defect that gene variation such as SNP cannot be detected are overcome.

The application belongs to the technical field of biological sequence data processing, and discloses a method and system for predicting and deducing folding paths of RNA secondary structure, which comprises: inputting a to-be-predicted RNA sequence into a deep learning pre-training model to perform feature extraction and pattern recognition, and obtaining a plurality of local structure modules each composed of at least three nucleotide pairs; regarding all the local structure modules as an action space of a reinforcement learning agent, and enabling the reinforcement learning agent to perform filtering and pairing actions on the local structure modules, gradually output a prediction result of the RNA secondary structure, and record the local structure modules and pairing sequences selected at each step to obtain a folding path, wherein the reinforcement learning agent adopts a composite reward function comprising a reward structure module formation, a punishment topological conflict, and an evaluation of local stability, and adaptively adjusts the weights of the composite reward function according to a folding stage and a current state. The application can improve the calculation efficiency and generalization ability, and reproduce the RNA folding path.

Provided herein are technologies for framing and evaluating biological sequence-based analysis tasks in a unified, natural-language-based, text in and text out format. Among other things, methods and systems of the present disclosure provide machine-learning technologies for combining biological sequence data, representing, for example, DNA, RNA, and protein sequences, with natural language, conversational style prompts that set out particular analysis tasks to be performed on the biological sequence data. This approach, for example, allows complex analysis tasks, including, but not limited to, identification of various sequence modifications, genes, and regulatory elements in DNA sequences, and quantification of properties such as degradation propensity of RNA and protein stability, to be input to a machine learning model in a uniform text-based format and for output to be generated in a same, unified, text-based format.

The present disclosure provides a siRNA silencing efficiency prediction method, device, electronic equipment and program product, comprising the following steps: obtaining a pseudo-label feature value of a to-be-predicted siRNA sequence based on a language model; searching for an off-target effect score of the to-be-predicted siRNA sequence based on a search model; obtaining a rule score of the to-be-predicted siRNA sequence based on a preset rule; performing feature extraction based on the to-be-predicted siRNA sequence to obtain sequence features; and obtaining a silencing efficiency of the to-be-predicted siRNA sequence based on a machine learning model according to the pseudo-label feature value, the off-target effect score, the rule score and the sequence features. The present disclosure comprehensively utilizes the biological sequence features of traditional rules and the strong learning ability of deep learning models, improves the accuracy and robustness of siRNA silencing efficiency prediction, reduces the dependence on a large amount of labeled data, and enhances the generalization ability among different biological samples, thereby providing a more reliable and practical tool for siRNA design.

The present disclosure provides methods and systems for building generative biological sequence models and synthesizing samples from those models at petascale. A library of nucleic acids or peptides can be designed in silico and a library can be synthesized in vitro.

The application discloses an anticancer peptide generation and screening method and system based on a diffusion model and a seed synchronous autoencoder, and comprises the following steps: preprocessing original anticancer peptide and non-anticancer peptide sequences and extracting multi-dimensional physicochemical features; training a diffusion model in a continuous numerical space based on sequence mapping based on a global random seed, adding noise through forward diffusion and learning the potential distribution of anticancer peptide data features through reverse denoising; based on a reverse seed synchronization strategy, training an autoencoder using the same random seed and environment state, and establishing a mapping from continuous random noise to discrete biological sequences; using the diffusion model to sample and generate potential vectors in the latent space and mapping them into candidate anticancer peptide sequences through the decoded synchronously trained decoder; inputting an integrated learning classifier based on grouping features, selecting the optimal feature group and predicting the anticancer activity probability, and outputting the anticancer characteristic peptide chain; the application solves the mapping contradiction between continuous space and discrete sequences and can efficiently screen anticancer peptides with high activity sequences.

The invention belongs to the technical field of biological research, and particularly relates to a biological sequence function prediction method and system. The system comprises: an input module for inputting a sequence of at least one biological sequence to be predicted; the data preprocessing module is configured to combine a biological sequence to be predicted with a fixed vocabulary comprising inverse document frequency (IDF) values of lexical items, and process the biological sequence to be predicted through a TF-IDF method to obtain TF-IDF vector representation of the biological sequence to be predicted; and the prediction module is used for inputting the TF-IDF vector representation into an artificial intelligence model to obtain a prediction result about whether the biological sequence to be predicted has a specific function or not. The invention also provides a method for predicting the function of the biological sequence by using the system. According to the method, the TF-IDF method is applied to characterization of the biological sequence for the first time, the characterization result and the machine learning technology are combined to be used for predicting the relation between the specific sequence and the biological function, and the method has good application prospects.

The invention provides a model training method, a biological sequence class task processing method and a related device, and the model training method comprises the steps: obtaining a first data set, and training a first neural network based on the first data set, thereby obtaining a biological sequence base model. Wherein the first data set comprises biological sequences of multiple sequence types. The biological sequences in the first data set have corresponding first sequence type markers and first sequence element markers, the first sequence type markers are used for indicating sequence types to which the corresponding biological sequences belong, and the first sequence element markers are used for indicating fragment types to which constituent fragments in the corresponding biological sequences belong. The biological sequence base model is used for constructing a sequence task model for executing a plurality of biological sequence tasks. According to the scheme, the biological sequence base model with good universality can be obtained, and the method can adapt to various different biological sequence tasks.

The present disclosure provides methods and systems that can ascertain how genetic variants impact molecular phenotypes. Such methods and systems may use additional conservation information. In an aspect, the present disclosure provides a method for training a molecular phenotype neural network (MPNN), comprising: (a) providing a molecular phenotype neural network (MPNN) comprising one or more parameters; (b) providing a training data set comprising (i) a set of one or more inputs comprising biological sequences and (ii) for each input in the set of one or more inputs, a set of one or more molecular phenotypes corresponding to the input; (c) configuring the one or more parameters of the MPNN based on the training data set to minimize a total loss of the training data set, thereby training the MPNN; and (d) outputting the one or more parameters of the MPNN.

An apparatus includes a memory configured to store a sequence. The sequence includes an estimation of a biological sequence. The sequence includes a set of elements. The apparatus also includes a set of hardware processors. Each hardware processor is configured to implement a segment processing module. The apparatus also includes an assignment module implemented in a hardware processor. The assignment module is configured to receive the sequence from the memory, and assign each element to at least one segment from a set of segments, including, when an element maps to at least a first segment and a second segment, assigning the element to both the first segment and the second segment. The segment processing module is configured to, for each segment from a set of segments specific to that hardware processor, and substantially simultaneous with the remaining hardware processors, remove at least a portion of duplicate elements in that segment to generate a deduplicated segment. The segment processing module is further configured to reorder the elements in the deduplicated segment to generate a realigned segment that has a reduced likelihood for alignment errors.

The invention relates to a computer-implemented method for creating a searchable indexed data structure for biological sequence data. The invention further relates to searching in said data structure for any given biological sequence data or associated information. The invention relates to a computer-implemented method comprising: a. providing biological sequence input data, b. decomposing the biological input data into non-overlapping portions, c. indexing each portion of the biological input data with at least one associated information of said portion comprised within or provided with the biological input data, thereby creating a two-tiered index for each portion. The method may further comprise d. receiving at least one query from a user, wherein the query comprises biological sequence information, language-based prompts and / or queries regarding the metadata comprised within or provided with the biological input data. The method may further comprise e. providing at least one output comprising at least one biological sequence information, biological sequence identifier and / or metadata of one or more portions of the biological sequence input data corresponding to the query. The invention further relates to the use of the method of the invention for identifying tissue and / or cell localized expression of at least one nucleic acid, such as a gene, and / or protein of interest. The method further relates to a data structure and a computer-readable medium comprising a data structure, such as a database, constructed according to the invention.

To provide a method, a computer program and a computer-implemented method for classifying biological sequences (identification of unknown genomes and closest known genomes).SOLUTION: A method for classifying biological sequences comprises: providing a training set related to a biological sequence; dividing the training set and the biological sequence into sequence fragments; extracting feature vectors from the sequence fragments of the training set and from the sequence fragments of the biological sequence; identifying feature vectors extracted from the sequence fragments of the biological sequence that correspond to feature vectors extracted from the sequence fragments of the training set; establishing a threshold value comprising an average degree of divergence of the feature vectors from the sequence fragments of the biological sequence, from the feature vectors from the sequence fragments of the training set; and detecting an anomaly among the extracted features of the sequence fragments of the biological sequence.SELECTED DRAWING: None

The invention discloses a method for constructing a phylogenetic tree for excavating antibiotic resistance based on microbial mutation sites, and relates to the technical field of microorganisms. The method comprises the following steps: step 1, database retrieval: determining a database according to a retrieval demand, analyzing a retrieval subject, extracting keywords, constructing a retrieval formula, and performing retrieval in the database to obtain drug resistance information and gene locus mutation data of microorganisms; 2, data collection and arrangement: screening and analyzing useful information from retrieval results, collecting research results about classification, action mechanisms, drug resistance conditions and the like of different antibiotics, creating a worksheet in Excel, and recording information such as microorganism names, strains, drug resistance conditions, drug resistance site mutation and the like; according to the application, a microorganism 23rRNA sequence is utilized to operate in a plurality of databases and software, and the method comprises the following steps: comparing homologous genes in an NCBI (National Center of Biotechnology Information) sub-website, constructing an evolutionary tree, retrieving drug resistance information in a plurality of professional websites, and counting and classifying antibiotics in Excel.