Jul 2026· Annual International Computer Software and Applications Conference· pp. 3022-3027· 0 citations· 13 references
Abstract
Small non-coding RNAs (ncRNAs) are functionally diverse RNA molecules involved in various regulatory and cellular processes. Accurate family classification of ncRNAs is important for large-scale functional annotation. However, many existing computational approaches rely heavily on predicted secondary structural information, which may not always be available. In this work, we propose a purely sequence-based deep learning framework for small ncRNA family classification without using explicit secondary structure features. Our approach combines one-hot encoded sequences, sequence-level motifs, normalized length, GC content, and biologically motivated handcrafted descriptors within a DenseNet-inspired convolutional neural network architecture. The proposed model was evaluated on the benchmark nRC dataset containing 13 ncRNA families using 10fold cross-validation and an independent test set. Our method achieved an MCC of 0.951 and an F1-score of 0.954, outperforming existing sequence-based as well as several structure-aware methods. The model also demonstrated improved performance for challenging classes such as miRNA, CD-box, HACA-box, and riboswitches. In addition, we evaluated the generalization capability of our framework on the recently introduced NCY dataset, where it maintained strong and consistent classification performance. These results show that enriched sequence-derived representations can effectively capture discriminative functional signals for ncRNA family classification without relying on secondary structure prediction. The source code is available at https://github.com/sharma-nandita/small_RNA_classification.
Identifying transcription factor binding sites (TFBSs) is fundamental to understanding complex gene regulatory mechanisms and the functions of non-coding regions. Although existing methods have achieved substantial strides, capturing both local structural features and long-range spatial dependencies within DNA sequences remains a major challenge for improving prediction accuracy. In this study, we propose DNCLA, a deep learning model that synergizes multisize convolutional fusion, Bidirectional Long ShortTerm Memory (Bi-LSTM) networks, and a multi-head self-attention
mechanism. At the feature extraction level, DNCLA breaks through the limitations of traditional single-sequence encoding by fusing Nucleotide Chemical Properties (NCP) with Dinucleotide Physicochemical Properties (DPCP). NCP provides a refined characterization of chemical differences between bases based on ring structures, hydrogen bond sites, and functional group properties, while DPCP introduces parameters such as local structural stability and geometric flexibility of the DNA. Subsequently, the model extracts spatial evolution from these high-dimensional features through a multi-size convolutional module; captures long-range spatial dependencies using Bi-LSTM layers; and employs a multi-head self-attention mechanism to achieve adaptive weight distribution of global features, thereby enhancing the perception of key regulatory motifs. Results from training and testing the proposed model on 165 ChIPseq datasets demonstrate that DNCLA possesses robust generalization capabilities and high predictive performance in TFBSs identification. This suggests that the incorporation of physicochemical features better elucidates the essence of interactions between transcription factors and DNA.
Jingjue Wei, Jie Feng· Match-communications in Math...· 0 citations
RNA plays a pivotal role in diverse cellular processes, and the rational design of functional RNA sequences is central to advancing RNA engineering. While deep generative models have shown significant promise, their generated sequences frequently lack the structural accuracy and evolutionary fidelity required for biological functionality. To address this challenge, we introduced VAERNAGen, a novel variational autoencoder-based framework for de novo generation of RNA family sequences. VAERNAGen's core innovation is a joint representation that encode aligned nucleotide sequences and their secondary structures into an 11-channel L × L two-dimensional matrix. This image-like format enables 2D convolutional neural networks to effectively learn the spatial interplay between sequence conservation and structural variation. Evaluated on two canonical Rfam families-RF00001 (5S ribosomal RNA) and RF00005 (tRNA), VAERNAGen outperformed the current state-of-the-art grammar-based method, achieving significantly higher median bit scores (105.25 vs. 89.34 for RF00001; 58.24 vs. 50.02 for RF00005). Generated sequences also exhibited greater nucleotide-level similarity to natural seed alignments and lower variance, reflecting enhanced biological plausibility and reproducibility. Together, these results establish VAERNAGen as a state-of-the-art method for de novo generation of RNA family sequences.
Guohua Huang, Yuyan Li, Changdong Yu et al.· International Journal of Bio...· 0 citations
A concatenated deep neural network that incorporates both peptide and nucleotide embedding from masked language modeling (MLM), being able to discover complex associations between sequence information and lncRNA classification.
This study evaluates four encoding schemes—one-hot, k-mer (substring-based encoding), embeddings, and Position-Specific Scoring Matrix (PSSM) using Convolutional Neural Networks (CNNs) and Long Short-Term Memories (LSTMs) and shows that k-mer encoding achieved the highest accuracy.
T. Kurniawan, Deshinta Arova Dewi, Randy Joy Magno Ventayen· Journal of Advances in Infor...· 0 citations
Structural variants are a major source of genomic variation and contribute to human disease and evolution through diverse mechanisms, yet their functional interpretation remains challenging. We present CADD-SV v2.0, an improved machine learning framework for scoring SV deleteriousness that expands on the original CADD-SV implementation. This version introduces a unified Random Forest model trained on an expanded set of proxy-neutral and proxy-deleterious variants drawn from human and non-human primate genomes. The model integrates updated genomic annotations, including constraint metrics, regulatory elements, and chromatin architecture features. It scores Deletions, Insertions, Duplications and Inversions based on a single scoring framework that uses both the variant and its flanking regions. To complement this framework, we also explore sequence-based annotations derived from SegmentNT, a deep learning model that provides functional predictions from DNA sequence at nucleotide resolution. Our analysis evaluated whether sequence-derived functional signals can provide additional information for SV prioritization and whether additional models with these features alone or in combination with previous coordinate-based annotations can be used.\ CADD-SV v2.0 outperforms its previous version and other tools in prioritizing deleterious variants across major SV types, including some previously unsupported, and substantially improves the computational workflow, increasing predictive power for genome-wide SV interpretation.
The novel combination of LZ complexity–based negative sample selection, CT feature representation, and GA-optimized CNN–LSTM architecture provides a robust and biologically informed framework for PPI prediction.