High-Order Convolutional Neural Network Architecture for Predicting DNA-Protein Binding Sites

Capsule Network for Predicting RNA-Protein Binding Preferences Using Hybrid Feature

RNA-Protein binding is involved in many different biological processes. With the progress of technology, more and more data are available for research. Based on these data, many prediction methods have been proposed to predict RNA-Protein binding preference. Some of these methods use only RNA sequence features for prediction, and some methods use multiple features for prediction. But, the performance of these methods is not satisfactory. In this study, we propose an improved capsule network to predict RNA-protein binding preferences, which can use both RNA sequence features and structure features. Experimental results show that our proposed method iCapsule performs better than three baseline methods in this field. We used both RNA sequence features and structure features in the model, so we tested the effect of primary capsule layer changes on model performance. In addition, we also studied the impact of model structure on model performance by performing our proposed method with different number of convolution layers and different kernel sizes.

AC-Caps: Attention Based Capsule Network for Predicting RBP Binding Sites of LncRNA

Long non-coding RNA(lncRNA) is one of the non-coding RNAs longer than 200 nucleotides and it has no protein encoding function. LncRNA plays a key role in many biological processes. Studying the RNA-binding protein (RBP) binding sites on the lncRNA chain helps to reveal epigenetic and post-transcriptional mechanisms, to explore the physiological and pathological processes of cancer, and to discover new therapeutic breakthroughs. To improve the recognition rate of RBP binding sites and reduce the experimental time and cost, many calculation methods based on domain knowledge to predict RBP binding sites have emerged. However, these prediction methods are independent of nucleotides and do not take into account nucleotide statistics. In this paper, we use a high-order statistical-based encoding scheme, then the encoded lncRNA sequences are fed into a hybrid deep learning architecture named AC-Caps. It consists of a joint processing layer(composed of attention mechanism and convolutional neural network) and a capsule network. The AC-Caps model was evaluated using 31 independent experimental data sets from 12 lncRNA-binding proteins. In experiments, our method achieves excellent performance, with an average area under the curve (AUC) of 0.967 and an average accuracy (ACC) of 92.5%, which are 0.014, 2.3%, 0.261, 28.9%, 0.189, and 21.8% higher than HOCCNNLB, iDeepS, and DeepBind, respectively. The results show that the AC-Caps method can reliably process the large-scale RBP binding site data on the lncRNA chain, and the prediction performance is better than existing deep-learning models. The source code of AC-Caps and the datasets used in this paper are available at https://github.com/JinmiaoS/AC-Caps .

LPI-CNNCP: Prediction of lncRNA-protein interactions by using convolutional neural network with the copy-padding trick

Long noncoding RNAs (lncRNAs) play critical roles in many pathological and biological processes, such as post-transcription, cell differentiation and gene regulation. Increasingly more studies have shown that lncRNAs function through mainly interactions with specific RNA binding proteins (RBPs). However, experimental identification of potential lncRNA-protein interactions is costly and time-consuming. In this work, we propose a novel convolutional neural network-based method with the copy-padding trick (named LPI-CNNCP) to predict lncRNA-protein interactions. The copy-padding trick of the LPI-CNNCP convert the protein/RNA sequences with variable-length into the fixed-length sequences, thus enabling the construction of the CNN model. A high-order one-hot encoding is also applied to transform the protein/RNA sequences into image-like inputs for capturing the dependencies among amino acids (or nucleotides). In the end, these encoded protein/RNA sequences are feed into a CNN to predict the lncRNA-protein interactions. Compared with other state-of-the-art methods in 10-fold cross-validation (10CV) test, LPI-CNNCP shows the best performance. Results in the independent test demonstrate that our LPI-CNNCP can effectively predict the potential lncRNA-protein interactions. We also compared the copy-padding trick with two other existing tricks (i.e., zero-padding and cropping), and the results show that our copy-padding rick outperforms the zero-padding and cropping tricks on predicting lncRNA-protein interactions. The source code of LPI-CNNCP and the datasets used in this work are available at https://github.com/NWPU-903PR/LPI-CNNCP for academic users.

Adaptive multi-source multi-view latent feature learning for inferring potential disease-associated miRNAs

Accumulating evidence has shown that microRNAs (miRNAs) play crucial roles in different biological processes, and their mutations and dysregulations have been proved to contribute to tumorigenesis. In silico identification of disease-associated miRNAs is a cost-effective strategy to discover those most promising biomarkers for disease diagnosis and treatment. The increasing available omics data sources provide unprecedented opportunities to decipher the underlying relationships between miRNAs and diseases by computational models. However, most existing methods are biased towards a single representation of miRNAs or diseases and are also not capable of discovering unobserved associations for new miRNAs or diseases without association information. In this study, we present a novel computational method with adaptive multi-source multi-view latent feature learning (M2LFL) to infer potential disease-associated miRNAs. First, we adopt multiple data sources to obtain similarity profiles and capture different latent features according to the geometric characteristic of miRNA and disease spaces. Then, the multi-modal latent features are projected to a common subspace to discover unobserved miRNA-disease associations in both miRNA and disease views, and an adaptive joint graph regularization term is developed to preserve the intrinsic manifold structures of multiple similarity profiles. Meanwhile, the Lp,q-norms are imposed into the projection matrices to ensure the sparsity and improve interpretability. The experimental results confirm the superior performance of our proposed method in screening reliable candidate disease miRNAs, which suggests that M2LFL could be an efficient tool to discover diagnostic biomarkers for guiding laborious clinical trials.

A classification model for lncRNA and mRNA based on k-mers and a convolutional neural network

Background

Long-chain non-coding RNA (lncRNA) is closely related to many biological activities. Since its sequence structure is similar to that of messenger RNA (mRNA), it is difficult to distinguish between the two based only on sequence biometrics. Therefore, it is particularly important to construct a model that can effectively identify lncRNA and mRNA.

Results

First, the difference in the k-mer frequency distribution between lncRNA and mRNA sequences is considered in this paper, and they are transformed into the k-mer frequency matrix. Moreover, k-mers with more species are screened by relative entropy. The classification model of the lncRNA and mRNA sequences is then proposed by inputting the k-mer frequency matrix and training the convolutional neural network. Finally, the optimal k-mer combination of the classification model is determined and compared with other machine learning methods in humans, mice and chickens. The results indicate that the proposed model has the highest classification accuracy. Furthermore, the recognition ability of this model is verified to a single sequence.

Conclusion

We established a classification model for lncRNA and mRNA based on k-mers and the convolutional neural network. The classification accuracy of the model with 1-mers, 2-mers and 3-mers was the highest, with an accuracy of 0.9872 in humans, 0.8797 in mice and 0.9963 in chickens, which is better than those of the random forest, logistic regression, decision tree and support vector machine.

Modeling in-vivo protein-DNA binding by combining multiple-instance learning with a hybrid deep neural network

Modeling in-vivo protein-DNA binding is not only fundamental for further understanding of the regulatory mechanisms, but also a challenging task in computational biology. Deep-learning based methods have succeed in modeling in-vivo protein-DNA binding, but they often (1) follow the fully supervised learning framework and overlook the weakly supervised information of genomic sequences that a bound DNA sequence may has multiple TFBS(s), and, (2) use one-hot encoding to encode DNA sequences and ignore the dependencies among nucleotides. In this paper, we propose a weakly supervised framework, which combines multiple-instance learning with a hybrid deep neural network and uses k-mer encoding to transform DNA sequences, for modeling in-vivo protein-DNA binding. Firstly, this framework segments sequences into multiple overlapping instances using a sliding window, and then encodes all instances into image-like inputs of high-order dependencies using k-mer encoding. Secondly, it separately computes a score for all instances in the same bag using a hybrid deep neural network that integrates convolutional and recurrent neural networks. Finally, it integrates the predicted values of all instances as the final prediction of this bag using the Noisy-and method. The experimental results on in-vivo datasets demonstrate the superior performance of the proposed framework. In addition, we also explore the performance of the proposed framework when using k-mer encoding, and demonstrate the performance of the Noisy-and method by comparing it with other fusion methods, and find that adding recurrent layers can improve the performance of the proposed framework.