Table 1.

Summary of investigations employing deep learning approaches in lncRNA studies between 2021 and 2023.

Research Topics	Deep Learning Approaches
Prediction of lncRNA–disease associations	ACLDA, combining autoencoders, CNN, and attention mechanism [27]; CapsNet–LDA, predicting lncRNA–disease associations using capsule network and attention [28]; DBNLDA, deep belief network-based lncRNA–disease association prediction [29]; Deep learning cluster analysis of lncRNAs in heart failure [30]; DeepMNE, deep multi-network embedding for lncRNA–disease prediction [31]; DHNLDA, deep hierarchical network with stacked autoencoder and ResNet [32]; DMFLDA, deep matrix factorization for predicting lncRNA–disease associations [33]; Dual attention network, enhances the learning of lncRNA–disease feature sets [34]; GCRFLDA, graph convolutional matrix completion-based lncRNA–disease prediction [35]; gGATLDA, lncRNA–disease associations prediction via graph-level attention networks [36]; GSMV, learning of global dependencies and multi-semantics within heterogeneous graphs [37]; GTAN, graph neural network for predicting lncRNA–disease associations [38]; HGATLDA, heterogeneous graph attention network for lncRNA–disease associations [39]; HGNNLDA, heterogeneous graph neural network for lncRNA–disease association [40]; Identifying cancer transcriptome signatures via deep learning interpretation [41]; iLncRNAdis–FB, CNN with fusing biological feature blocks [42]; LDACE, combining extreme learning machine with CNN [43]; LDICDL, identifying lncRNA–disease associations using collaborative deep learning [44]; LGDLDA, predicting disease-related lncRNAs via multiomics data and machine learning [45]; LR–GNN, graph neural network-based prediction of molecular associations [46]; MAGCNSE, lncRNA–disease association prediction via multi-view graph convolutional network [47]; MCA–Net, predicting lncRNA–disease associations using attention CNN [48]; MLMKDNN, predicting ncRNA–disease associations via deep multiple kernel learning [49]; MLGCNET, predicting lncRNA–disease associations using multi-layer graph embedding [50]; Multi-run concrete autoencoder identifying prognostic lncRNAs for cancers [51]; NELDA, predicting lncRNA–disease associations via deep autoencoder models [52]; Novel computational approach, lncRNA–disease prediction via BPSO and ML–ELM [53]; PANDA, graph convolutional autoencoders predicting novel lncRNA–disease associations [54]; Prognostic and diagnostic value of lncRNA in colorectal cancer [55]; VADLP, predicting lncRNA–disease associations with attentional multi-level encoding [56]; VGAELDA, predicting lncRNA–disease associations using variational inference and autoencoders [57].
Prediction of lncRNA–protein interactions	BiHo–GNN, using bipartite graph embedding [58]; Capsule–LPI, a multichannel capsule network for lncRNA–protein interaction prediction [59]; DeepLPI, a multimodal deep learning method for lncRNA–protein isoform interactions [60]; DFRPI, deep autoencoder and marginal Fisher analysis [61]; EnANNDeep, ensemble-based framework with adaptive k-nearest neighbor for the lncRNA–protein interaction [62]; iEssLnc, graph neural network-based estimation of lncRNA gene essentiality [63]; LGFC–CNN, using deep learning with feature combination [64]; LPI–CSFFR, CNN-based lncRNA–protein interaction prediction with serial fusion and feature reuse [65]; LPI–deepGBDT, gradient boosting decision trees-based lncRNA–protein interaction identification [66]; LPI–DLDN, dual-net neural architecture for lncRNA–protein interactions prediction [67]; LPI–HyADBS, hybrid framework with DNN, XGBoost, SVM for lncRNA–protein interaction [68]; LPICGAE, predicting lncRNA–protein interactions using combined graph autoencoders [69]; PRPI–SC, ensemble deep learning for plant lncRNA–protein interactions prediction [70]; RLF–LPI, ensemble learning framework with residual LSTM and fusion attention [71].
Prediction of lncRNA–miRNA interactions	BoT–Net, efficient lncRNA–miRNA interaction prediction using the bag of tricks-based neural network [72]; DeepWalk–LMI, inferring lncRNA–miRNA associations via comprehensive graph [73]; GCNCRF, predicting lncRNA–miRNA interactions using graph convolution and conditional random field [74]; MD–MLI, predicting lncRNA–miRNA interactions using multiple features and hierarchical deep learning [75]; ncRNAInter, a novel strategy using a graph neural network to discover lncRNA–miRNA interactions [76]; Optimized ensemble deep learning, predicting plant lncRNA–miRNA based on artificial gorilla troops algorithm [77]; PmliHFM, plant lncRNA–miRNA interaction prediction via hybrid feature mining network [78]; PmliPEMG, multi-level information enhancement and greedy fuzzy decision for plant lncRNA–miRNA interaction prediction [79]; preMLI, uncovering potential lncRNA–miRNA interactions through pre-training and deep feature mining [80].
Classification and Prediction of lncRNA characteristics	Class similarity network, identifying lncRNAs using relationships among samples [81]; DeepLncPro, CNN for identifying lncRNA promoters [82]; DeepPlnc, high accuracy plant lncRNA identification using bimodal CNN [83]; Genome-wide analysis, exploring features related to human lncRNA stability [84]; LncDLSM, lncRNA identification using the deep learning-based sequence model [85]; LncReader, identifying dual-functional lncRNAs using multi-head self-attention [86]; lncIBTP, predicting interaction biomolecule type for a given lncRNA using ensemble deep learning [87]; RNA prediction based on neural network integration of CNN and Bi-LSTM [88]; Xlnc1DCNN, interpretable deep learning model, lncRNA identification using 1D CNN [89].
Prediction of lncRNA subcellular localization	DeepLncLoc, a deep learning framework for lncRNA subcellular localization using subsequence embedding [90]; EVlncRNA–Dpred, an improved prediction method of experimentally validated lncRNAs using deep learning [91]; GM–lncLoc, lncRNA subcellular localization prediction based on graph neural network with meta-learning [92]; GraphLncLoc, predicting lncRNA subcellular localization using graph convolutional networks and sequence-to-graph transformation [93]; PlncRNA–HDeep, a plant long non-coding RNA prediction method that utilizes hybrid deep learning with two encoding styles [94].
Prediction of functional roles of lncRNAs in immune response pathways	CD8–Net, constructing ceRNA networks for CD8 T cells in breast cancer [95]; JD–lncRNA–ID, identifying lncRNA associated with bovine Johne’s disease using neural networks and logistic regression [96];
Deep learning applications through the utilization of lncRNA input data	MVMTMDA, predicting miRNA–disease associations through lncRNA–miRNA interactions [97]; Predicting microsatellite instability in colorectal cancer using multimodal deep learning [98]; Predicting miRNA–disease associations, a method based on lncRNA–miRNA interactions and graph convolution networks [99]; WGAN–psoNN, tumor lymph node metastasis prediction using WGAN and psoNN [100].
Identification of lncRNA–protein-coding gene (PCG) associations	GAE-LGA, deep learning prediction of lncRNA–PCG associations with cross-omics correlation learning [101].