NMCMDA: neural multicategory MiRNA–disease association prediction

Abstract

Motivation There is growing evidence showing that the dysregulations of miRNAs cause diseases through various kinds of the underlying mechanism. Thus, predicting the multiple-category associations between microRNAs (miRNAs) and diseases plays an important role in investigating the roles of miRNAs in diseases. Moreover, in contrast with traditional biological experiments which are time-consuming and expensive, computational approaches for the prediction of multicategory miRNA–disease associations are time-saving and cost-effective that are highly desired for us. Results We present a novel data-driven end-to-end learning-based method of neural multiple-category miRNA–disease association prediction (NMCMDA) for predicting multiple-category miRNA–disease associations. The NMCMDA has two main components: (i) encoder operates directly on the miRNA–disease heterogeneous network and leverages Graph Neural Network to learn miRNA and disease latent representations, respectively. (ii) Decoder yields miRNA–disease association scores with the learned latent representations as input. Various kinds of encoders and decoders are proposed for NMCMDA. Finally, the NMCMDA with the encoder of Relational Graph Convolutional Network and the neural multirelational decoder (NMR-RGCN) achieves the best prediction performance. We compared the NMCMDA with other baselines on three experimental datasets. The experimental results show that the NMR-RGCN is significantly superior to the state-of-the-art method TDRC in terms of Top-1 precision, Top-1 Recall, and Top-1 F1. Additionally, case studies are provided for two high-risk human diseases (namely, breast cancer and lung cancer) and we also provide the prediction and validation of top-10 miRNA–disease-category associations based on all known data of HMDD v3.2, which further validate the effectiveness and feasibility of the proposed method.

HOPEXGB: A Consensual Model for Predicting miRNA/lncRNA-Disease Associations Using a Heterogeneous Disease-miRNA-lncRNA Information Network

“…Since HOPEXGB is a consensual model and cannot be directly compared to other algorithms, we extracted the lncRNA and miRNA prediction results from the final HOPEXGB model and gave the predictions for lncRNAs and miRNAs, respectively. Then we compared HOPEXGB with four other miRNA-disease association prediction models, including NMCMDA, 27 HGANMDA, 29 MDPBMP, 30 and DANE-MDA. 31 NMCMDA is an end-to-end learning-based method of multiple neural categories, and DANE-MDA is a deep learning method by using a deep stacked autoencoder; but both methods only integrated the information of miRNAs and diseases, lacking the interaction information of miRNAs with lncRNAs.…”

“…25 In addition, Zhou et al 26 have used the HOPE embedding method to predict lncRNA-disease associations by integrating the associations among miRNAs, lncRNAs, diseases, as well as proteins and drugs. As for the predictions of disease-related miRNAs, the graph neural network-based machine-learning methods have also been developed; for example, Wang et al 27 presented a novel data-driven end-to-end learning-based method of neural multiple-category miRNA-disease association prediction, Yan et al 28 used graph neural networks and miRNA sequence features to predict deep-level miRNA-disease associations, Li et al 29 proposed a deep learning model based on a hierarchical graph attention network for predicting miRNA-disease associations, Yu et al 30 adopted meta-paths to extract features to predict miRNA-disease associations, and Ji et al 31 used a deep stacked autoencoder on the diverse orders of matrixes containing structure and attribute information to extract the integrated features and then trained them by using a random forest classifier to predict potential miRNA-disease associations. For these methods, how to efficiently and reliably compose a graph is a challenge, and the lack of interpretability of convolutional networks also restricts the scalability of the prediction models.…”

MDSCMF: Matrix Decomposition and Similarity-Constrained Matrix Factorization for miRNA–Disease Association Prediction

“…Finally, a fully connected neural network combined linear representations with nonlinear representations to generate the predicted miRNA–disease association scores. Wang et al [ 39 ] presented a novel method called NMCMDA to observe unknown disease-related miRNAs. The encoder and decoder were the two essential components in NMCMDA.…”

ILPMDA: Predicting miRNA–Disease Association Based on Improved Label Propagation

“…However, MLPMDA requires high-quality biological data to achieve reliable and stable performance. A novel method of neural multiple-category miRNA–disease association prediction named NMCMDA was proposed by Wang et al (2021) to observe the unknown disease-related miRNAs. The two main components in NMCMDA were encoder and decoder.…”