Abstract:Noncoding RNAs (ncRNAs) play crucial roles in many cellular life activities by interacting with proteins. Identification of ncRNA-protein interactions (ncRPIs) is key to understanding the function of ncRNAs. Although a number of computational methods for predicting ncRPIs have been developed, the problem of predicting ncRPIs remains challenging. It has always been the focus of ncRPIs research to select suitable feature extraction methods and develop a deep learning architecture with better recognition performance. In this work, we proposed an ensemble deep learning framework, RPI-EDLCN, based on a capsule network (CapsuleNet) to predict ncRPIs. In terms of feature input, we extracted the sequence features, secondary structure sequence features, motif information, and physicochemical properties of ncRNA/protein. The sequence and secondary structure sequence features of ncRNA/protein are encoded by the conjoint k-mer method and then input into an ensemble deep learning model based on CapsuleNet by combining the motif information and physicochemical properties. In this model, the encoding features are processed by convolution neural network (CNN), deep neural network (DNN), and stacked autoencoder (SAE). Then the advanced features obtained from the processing are input into the CapsuleNet for further feature learning. Compared with other state-of-the-art methods under 5-fold cross-validation, the performance of RPI-EDLCN is the best, and the accuracy of RPI-EDLCN on RPI1807, RPI2241, and NPInter v2.0 data sets was 93.8%, 88.2%, and 91.9%, respectively. The results of the independent test indicated that RPI-EDLCN can effectively predict potential ncRPIs in different organisms. In addition, RPI-EDLCN successfully predicted hub ncRNAs and proteins in <i>Mus musculus</i> ncRNA-protein networks. Overall, our model can be used as an effective tool to predict ncRPIs and provides some useful guidance for future biological studies.

SAWRPI: A Stacking Ensemble Framework With Adaptive Weight for Predicting ncRNA-Protein Interactions Using Sequence Information

RPI-SE: a Stacking Ensemble Learning Framework for Ncrna-Protein Interactions Prediction Using Sequence Information

RPI-MDLStack: Predicting RNA–protein interactions through deep learning with stacking strategy and LASSO

Efficient Framework For Predicting Ncrna-Protein Interactions Based On Sequence Information By Deep Learning

RPI-Pred: predicting ncRNA-protein interaction using sequence and structural information

A Deep Learning Framework for Robust and Accurate Prediction of Ncrna-Protein Interactions Using Evolutionary Information

Computational Methods for Predicting Ncrna-Protein Interactions

Learning Distributed Representations of RNA and Protein Sequences and Its Application for Predicting Lncrna-Protein Interactions

PRPI-SC: an ensemble deep learning model for predicting plant lncRNA-protein interactions

RPIPCM: A deep network model for predicting lncRNA-protein interaction based on sequence feature encoding

IRWNRLPI: Integrating Random Walk and Neighborhood Regularized Logistic Matrix Factorization for lncRNA-Protein Interaction Prediction

DEEPStack-RBP: Accurate identification of RNA-binding proteins based on autoencoder feature selection and deep stacking ensemble classifier

LPI-BLS: Predicting Lncrna–protein Interactions with a Broad Learning System-Based Stacked Ensemble Classifier

The Bipartite Network Projection-Recommended Algorithm for Predicting Long Non-coding RNA-Protein Interactions.

Prediction of Plant lncRNA-Protein Interactions Using Sequence Information Based on Deep Learning

RPI-EDLCN: An Ensemble Deep Learning Framework Based on Capsule Network for ncRNA–Protein Interaction Prediction

MHAM-NPI: Predicting ncRNA-protein interactions based on multi-head attention mechanism

Bgfe: A Deep Learning Model For Ncrna-Protein Interaction Predictions Based On Improved Sequence Information

IPMiner: hidden ncRNA-protein interaction sequential pattern mining with stacked autoencoder for accurate computational prediction

De Novo Prediction of RNA-protein Interactions from Sequence Information.

LncRNA-Protein Interaction Prediction Based on Regularized Nonnegative Matrix Factorization and Sequence Information