A Deep Learning-Based Identification Method for RNA and Protein Binding Sites

Abstract

The invention discloses a recognition method of RNA and protein binding sites based on deep learning. The method is as follows: the recognition method is divided into two parts: data processing and model design training, and the main steps of each part are as follows: the first step, data processing : Step 1, remove redundancy from original data; Step 2, predict secondary structure information; Step 3, encode the secondary structure obtained in Step 2; Step 4, save all data in the corresponding csv file as input data ;The second step, model design and training: step 1, coded into a data matrix as the model input; step 2, the dimensions of the two feature matrices are the same; step 3, input the combined features into the encoder; step 4, the final encoder encoding The result; Step 5, obtain the final category through the sigmoid function. Beneficial effects: the use of the Transformer network to learn the long-term dependence of sequences and encode features can effectively predict the binding sites of RNA and proteins.

Description

technical field [0001] The invention relates to a method for identifying binding sites of RNA and proteins, in particular to a method for identifying binding sites of RNA and proteins based on deep learning. Background technique [0002] RNA binding protein (RNA binding protein, RBP) is a kind of protein that regulates RNA metabolic process. It is closely related to many important biological processes, such as post-transcriptional regulation of genes and gene expression. For example, miRNA is a type of RNA with a length of about 21 bp. After interacting with mRNA, it will stop translation and play a role in post-transcriptional regulation. In addition, the disorder of RBP may lead to various diseases. Mutations of RBP, FUS, and TDP-43 are closely related to ALS. RBP can also disrupt the post-transcriptional mechanism of diabetes. Understanding the regulatory role of RBP in diabetes can help design RNA-based treatments for diabetic complications. Integrating the information...

AI Technical Summary

Problems solved by technology

At present, there are two main methods for studying the binding sites of RNA and proteins. One is through biological experiments. This method not only has strict requirements on the experimental environment, but also has high requirements on the professionalism of researchers.

Although this method is accurate and reliable, it requires a lot of time and effort, and many uncontrollable factors in the experimental process will also have a huge impact on the results

Images