CRBPDL: Identification of circRNA-RBP interaction sites using an ensemble neural network approach

Abstract

Circular RNAs (circRNAs) are non-coding RNAs with a special circular structure produced formed by the reverse splicing mechanism. Increasing evidence shows that circular RNAs can directly bind to RNA-binding proteins (RBP) and play an important role in a variety of biological activities. The interactions between circRNAs and RBPs are key to comprehending the mechanism of posttranscriptional regulation. Accurately identifying binding sites is very useful for analyzing interactions. In past research, some predictors on the basis of machine learning (ML) have been presented, but prediction accuracy still needs to be ameliorated. Therefore, we present a novel calculation model, CRBPDL, which uses an Adaboost integrated deep hierarchical network to identify the binding sites of circular RNA-RBP. CRBPDL combines five different feature encoding schemes to encode the original RNA sequence, uses deep multiscale residual networks (MSRN) and bidirectional gating recurrent units (BiGRUs) to effectively learn high-level feature representations, it is sufficient to extract local and global context information at the same time. Additionally, a self-attention mechanism is employed to train the robustness of the CRBPDL. Ultimately, the Adaboost algorithm is applied to integrate deep learning (DL) model to improve prediction performance and reliability of the model. To verify the usefulness of CRBPDL, we compared the efficiency with state-of-the-art methods on 37 circular RNA data sets and 31 linear RNA data sets. Moreover, results display that CRBPDL is capable of performing universal, reliable, and robust. The code and data sets are obtainable at https://github.com/nmt315320/CRBPDL.git.

Fig 1. The overall framework of CRBPDL.

(A) The workflow of the development and assessment process of CRBPDL. (B) The structure of the sig-CRBPDL framework, including the input layer, convolutional layers, merger layers, inception module, attention layers, fully connected layers and output layer.

Fig 2

(A) Comparison of model performance between different network depths visualized by box and fiddle charts. (B) Model performance analysis under different EPOCH. (C) Comparison of model performance under different learning rate schemes. (D) Comparison of model performance under different feature coding schemes.

Fig 3

(A) Performance comparison of five feature codes.(B) Heat maps of different network model performance under 37 data sets.(C) T -SNE scatter plot with original feature coding. (D) T-SNE Scatter Diagram of Deep Feature after Deep Convolutional Network.

Fig 4

(A) Performance comparison of MSRN and BiGRU. (B) The performance comparison between CRBPDL integrated model and various classification algorithms (C) ROC curves of 37 datasets under the integration model (D) Radar chart of ACC indicators of CRBPDL model under 31 lncRNA datasets.