10 results about "Gene selection" patented technology

The invention provides a feature gene selection method and system based on deep learning attribution analysis and beam combination optimization, and relates to the technical field of signal analysis, and the method comprises the following steps: obtaining single cell RNA sequencing data and batch RNA sequencing data; the method comprises the following steps: based on single-cell RNA sequencing data, obtaining an initial gene pool aiming at a plurality of disease subtypes through a consensus screening strategy fusing XGBoost multi-dimensional importance measurement and deep learning SHAP attribution analysis; performing first-stage optimization on the gene list of each subtype by taking the initial gene pool as a starting point and adopting beam combination search and based on a global discrimination objective function to obtain a first-stage optimized gene set; integrating the first-stage optimized gene set with the differential expression gene set of the batch RNA sequencing data, and carrying out second-stage optimization by adopting bundle combination search again to obtain a second-stage optimized gene set; and performing cross validation integration on the second-stage optimized gene set, and outputting a final feature gene set.

The present disclosure relates to providing an engineered CHO cell line wherein the two essential metabolic genes are knocked out. Particularly, the present invention relates to a double knockout CHO cell line (DHFR− / − and GS− / −) with disrupted Dihydrofolate Reductase (DHFR) and Glutamine Synthetase (GS) genomic loci. The double knockout CHO cell line (DHFR− / − and GS− / −) being suitable for expression of monoclonal antibodies, dimeric therapeutic proteins, Fab, single chain fragments, or the like. The present disclosure also provides method of generation of a double knock out CHO cell line (DHFR− / − and GS− / −) using gene selection and / or manipulating techniques such as CRISPR / Cas9 system, Zinc Finger Nuclease, TALEN, or the like. The present disclosure further provides method of selection of clones and production of therapeutic proteins of interest with increased titre.

The present application belongs to the tumor precision medicine and artificial intelligence medical technology field, and particularly relates to a non-small cell lung cancer neoadjuvant immunotherapy efficacy prediction method based on transfer learning. The method first pretreats the non-small cell lung cancer immunotherapy cohort and the neoadjuvant immunotherapy cohort, selects immunotherapy and neoadjuvant immunotherapy characteristic genes, and based thereon, carries out deep learning training on the immunotherapy cohort to obtain a basic model, finally, the basic model is fine-tuned to the neoadjuvant immunotherapy training data set through transfer learning, and finally used to the neoadjuvant immunotherapy cohort to predict the neoadjuvant immunotherapy effect. The present application improves the stability and generalization ability of the model in the small sample neoadjuvant treatment scene through gene selection and transfer fine-tuning, and can be used for clinical auxiliary decision and population stratification.

The invention discloses an efficient SNP marker automatic screening system for dairy cow breeding, relates to the technical field of dairy cow breeding screening, and aims to solve the problems that the scientificity and stability of a screening result are reduced, and the error between genetic improvement efficiency and marker utilization rate is increased. The method comprises the following steps: extracting dairy cow breeding character information and gene site data from a genotype-phenotype database, calculating heritability of each character, screening target characters with high genetic contribution degree, calculating character action characteristics and effect directions of gene sites on the basis of an association relationship between the target characters and SNP sites, quantitatively describing regulation intensity of the gene sites, and screening out the target characters with high genetic contribution degree. Generating a character guiding coefficient according to the character guiding coefficient, screening significant candidate sites in combination with a threshold value, performing cross-character frequency statistics on the candidate sites, calculating a breeding efficiency index in combination with a character correlation weight, and performing hierarchical marking according to the index, so as to realize multi-character comprehensive genetic effect evaluation and automatic screening. The breeding data processing efficiency and the gene selection scientificity are improved.

The invention discloses a method for selecting single nucleotide polymorphism (SNP) for phenotype prediction based on feature importance, which comprises the following steps of: performing missing filling on SNP data, and dividing a sample into a training set and a test set before phenotype-related screening to avoid data leakage; then, by taking the gene as a unit, fitting phenotypes of the SNPs positioned on the promoter, the exon and the intron by adopting a regression model, calculating correlation coefficients and carrying out multiple inspection correction, and taking a significant correlation gene as a candidate; selecting one SNP (Single Nucleotide Polymorphism) from each gene in the candidate genes on the basis of feature importance to form an Important-SNP set; and encoding the set, and inputting the encoded set into a prediction model to obtain a phenotype prediction result. According to the method, the feature dimension is remarkably reduced while the prediction accuracy is maintained or improved, and the method has relatively high interpretability and engineering availability, is suitable for phenotype prediction of crops, can be used as a core marker for design and optimization of a breeding chip, and provides efficient and interpretable technical support for molecular breeding and genome selection.

The invention discloses a quantitative evaluation system and method for the inflammation level of cells cultured in vitro, and the system employs a modular design, and is composed of a data collection and preprocessing module, a feature gene selection module, a neural network modeling module, and an inflammation index calculation module. The evaluation method comprises the steps of data preparation, original data processing, feature gene screening, machine learning model screening, neural network model architecture, neural network weight analysis, inflammation index calculation and the like. The method has the advantages of high speed, high throughput, low cost, high accuracy, high mobility and high interpretability.

This invention relates to a method for increasing the yield of polyketide compounds by splitting a modular polyketide synthase gene. (1) Constructing a RedEx gene cassette for splitting the modular polyketide synthase gene; (2) Constructing a target plasmid carrying an amp-ccdB gene selection expression cassette; (3) Constructing a target plasmid carrying an amp-ccdB gene selection expression cassette. C DD-TGA-RBS-ATG- N (3) Target plasmid encoding the DD coding sequence; (4) Construction of recombinant strains; (5) Fermentation culture to obtain polyketide compounds. This invention is the first to discover that the biosynthetic efficiency of host strains carrying the split PKS gene is much higher than that of host strains carrying the natural PKS gene, which can achieve the purpose of increasing the yield of polyketide compounds. Using this method of splitting modular polyketide synthase genes, high-yield butenyl spinosad genetically engineered bacteria and high-yield avermectin genetically engineered bacteria were successfully constructed, which effectively increased the yield of butenyl spinosad and avermectin. The yield of butenyl spinosad increased by a maximum of 30 times, and the yield of avermectin increased by a maximum of 5.8 times.

The application discloses a gene selection method based on biological character expression of multi-target feature selection, which comprises the following steps: step 1, inputting biological gene samples; step 2, grouping features of the input gene samples and generating a sub-population for each group; step 3, independently evolving each sub-population; step 4, gradually merging all the sub-populations; and step 5, finally merging all the sub-populations into one population, searching on the original genes, finally obtaining the best partial genes on the original gene set in expression effect, and taking the partial genes as the finally selected genes. The application can screen a small number of important genes from a large number of gene features, thereby effectively improving the accuracy of gene selection of high-dimensional biological character expression, and is also beneficial to shortening a research period and reducing research cost.

This invention discloses a gene selection system based on homomorphic encryption technology, comprising: an input module for inputting data; a selection module for allocating the input data to corresponding prediction models and obtaining all information from the data and pedigree files in a one-step process; an execution prediction module for reconstructing the computation array according to the configuration information provided by the selection module to implement the algorithms for each stage of homomorphic encryption; a display module for storing the initial plaintext, intermediate ciphertext, and plaintext obtained after decryption of the homomorphic encryption algorithm in the execution prediction module; and a cloud output module for outputting the encrypted data to a cloud platform. This invention also develops a local client, HEcryptor, based on the homomorphic encryption algorithm to work with the HEGS platform. Users can encrypt data locally before uploading to ensure complete data security. The encrypted data can be used for both single-field and cross-field genome selection, meeting various user needs.