Method and system for 2om methylation modification site recognition based on feature selection and deep neural network
By employing feature selection and deep neural network methods, and utilizing ANOVA feature selection and multi-head self-attention mechanism to process RNA sequences, this approach solves the problems of high time consumption, high cost, and low accuracy in existing RNA 2OM modification detection technologies, achieving more efficient and accurate identification of RNA modification sites.
CN119207585BActive Publication Date: 2026-06-26HAINAN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HAINAN UNIV
- Filing Date
- 2024-09-26
- Publication Date
- 2026-06-26
Smart Images
Figure CN119207585B_ABST
Abstract
The present application relates to a kind of 2OM methylation modification site recognition method and system based on feature selection and deep neural network.The method includes: the encoding mode of RNA sequence is converted to obtain sequence feature;The dimension of sequence feature is reduced by ANOVA feature selection algorithm and statistical value is counted, feature selection is carried out based on statistical value;For the RNA sequence of modification site in C, input feature is input into first classifier and output classification result;For the RNA sequence of modification site in A, G, U, input feature is input into second classifier and output classification result.The input feature is reduced by ANOVA feature selection algorithm, and the feature is selected based on variance analysis, and different classifiers are used to process the data of 2OM site in different positions, which can improve the recognition accuracy of 2OM methylation modification site.
Need to check novelty before this filing date? Find Prior Art