Dual-function RNA has both protein-coding and non-coding functions during gene regulation and organismal development. In response to the problems of high false positive rate and limited feature extraction ability of existing recognition methods, a recognition method BiF-GCN based on graph convolutional neural network is proposed. This method enhances RNA sequence representation by introducing topological graph structures, combines sequence-level features with graph structural features, and uses the graph convolutional network model for prediction. Results show that BiF-GCN performed better than existing methods in several evaluation areas, achieving an accuracy of 0.898 4, Matthews correlation coefficient of 0.799 3, and F1-score of 0.894 3. It reduced the false positive rate and improved the classification performance, providing an effective approach for bifunctional RNA recognition. The study confirms that integrating sequence and graph structural features is an effective approach for identifying bifunctional RNAs.
为了更好地捕获序列的全局重要特征,BiF-GCN引入一种基于投影的注意力机制。该机制设置注意力头数为4,隐藏层维度为64,值向量( V )维度为32。与传统的多头注意力机制不同,传统方法中查询( Q )、键( K )和值具有相同维度,且输出通过线性投影进行融合,BiF-GCN将 V 的维度与 Q 和 K 解耦,并直接将多个注意力头的输出[14]进行拼接。其计算公式定义如下:
SAMPATHK, EPHRUSSIA. CncRNAs:RNAs with both coding and non-coding roles in development[J]. Development,2016,143(8):1234-1241.
[2]
LIUZ, BAIT, LIUB,et al. MulStack:An ensemble learning prediction model of multilabel mRNA subcellular localization[J]. Computers in Biology and Medicine,2024,175:108-289.
[3]
RANSOHOFFJ D, WEIY, KHAVARIP A. The functions and unique features of long intergenic non-coding RNA[J]. Nature Reviews Molecular Cell Biology,2018,19(3):143-157.
[4]
MATTICKJ S, AMARALP P, CARNINCIP,et al. Long non-coding RNAs:definitions,functions,challenges and recommendations[J]. Nature Reviews Molecular Cell Biology,2023,24(6):430-447.
[5]
CHENGJ, LINY, XUL,et al. ViRBase v3. 0:a virus and host ncRNA-associated interaction repository with increased coverage and annotation[J]. Nucleic Acids Research,2022,50(D1):D928-D933.
[6]
O’CONNORO, MCVEIGHT P. Increasing use of artificial intelligence in genomic medicine for cancer care-the promise and potential pitfalls[J]. BJC Reports,2025,3(1):20.
[7]
HUANGY, WANGJ, ZHAOY,et al. cncRNAdb:a manually curated resource of experimentally supported RNAs with both protein-coding and noncoding function[J]. Nucleic Acids Research,2021,49(D1):D65-D70.
[8]
ZHAOY, LIH, FANGS,et al. NONCODE 2016:an informative and valuable data source of long non-coding RNAs[J]. Nucleic Acids Research,2016,44(D1):D203-D208.
[9]
VOLDERSP-J, ANCKAERTJ, VERHEGGENK,et al. LNCipedia 5:towards a reference set of human long non-coding RNAs[J]. Nucleic acids research,2019,47(D1):D135-D139.
[10]
ZHANGX, WANGY, WEIQ,et al. DRBPPred-GAT:Accurate prediction of DNA-binding proteins and RNA-binding proteins based on graph multi-head attention network[J]. Knowledge-Based Systems,2024,285:111354.
[11]
POSTICG,TAV C, PLATONL,et al. IRSOM2:a web server for predicting bifunctional RNAs[J]. Nucleic Acids Research,2023,51(W1):W281-W288.
[12]
LIUT, ZOUB, HEM,et al. LncReader:identification of dual functional long noncoding RNAs using a multi-head self-attention mechanism[J]. Briefings in Bioinformatics,2023,24(1):bbac579.
[13]
LIM, ZHAOB, YINR,et al. GraphLncLoc:long non-coding RNA subcellular localization prediction using graph convolutional networks based on sequence to graph transformation[J]. Briefings in Bioinformatics,2023,24(1):bbac565.
[14]
VASWANIA, SHAZEERN, PARMARN,et al. Attention is all you need[C]//Advances in Neural Information Processing Systems. Red Hook:Curran Associates Incorporated, 2017:5998-6008.
[15]
DENGC, TANGY, ZHANGJ,et al. RNAGCN:RNA tertiary structure assessment with a graph convolutional network[J]. Chinese Physics B,2022,31(11):118702.
[16]
LIW, GODZIKA. Cd-hit:a fast program for clustering and comparing large sets of protein or nucleotide sequences[J]. Bioinformatics,2006,22(13):1658-1659.
[17]
RUSSELJ, PINILLA-REDONDOR, MAYO-MUñOZD,et al. CRISPRCasTyper:automated identification,annotation,and classification of CRISPR-Cas loci[J]. The CRISPR Journal,2020,3(6):462-469.
[18]
PASZKEA, GROSSS, MASSAF,et al. Pytorch:An imperative style,high-performance deep learning library[C]//Advances in Neural Information Processing Systems. Red Hook:Curran Associates Incorporated, 2019:8024-8035.
[19]
DEGENHARDTM F, DEGENHARDTH F, BHANDARIY R,et al. Determining structures of RNA conformers using AFM and deep neural networks[J]. Nature,2025,637(8048):1234-1243.
[20]
LIY H, XUJ Y, TAOL,et al. SVM-Prot 2016:a web-server for machine learning prediction of protein functional families from sequence irrespective of similarity[J]. PLOS ONE,2016,11(8):e0155290.
[21]
TOUWW G, BAYJANOVJ R, OVERMARSL,et al. Data mining in the life sciences with random forest:a walk in the park or lost in the jungle[J]. Briefings in Bioinformatics,2013,14(3):315-326.
[22]
HUSNAINM, MISSENM M S, MUMTAZS,et al. Visualization of high-dimensional data by pairwise fusion matrices using t-SNE[J]. Symmetry,2019,11(1):107.