在多細胞生物的基因組中，轉錄調控是通過轉錄因子（transcription factor, TF）或調控蛋白與特定的去氧核醣核酸（DNA）序列結合來實現基因表達調控的過程。這些特定的DNA序列被稱為順式調控模組（cis-regulatory module, CRM），而受調控的基因則被稱為目標基因（target gene）。本研究旨在開發一種全面性的工具，用於預測CRM是否調控特定基因。在以往的研究中，CRM的目標基因主要基於CRM與基因啟動子之間的相互作用來確定。然而，最近的研究表明，加強子可以通過加強子RNA（eRNAs）的轉錄來遠距離目標基因，而不僅僅是通過加強子與啟動子的相互作用。同時發現，CRM和基因之間的距離並不能直接解釋它們的調控關係。例如，CRM可能與其目標基因的啟動子近側區域，且基因可能不受其最近的CRM調控。本研究開發了兩個模型，分別預測CRM和基因的交互作用以及CRM和基因在機制RNA上的調控關係，從而綜合考慮交互作用和加強子RNA的調控機制。本研究要求：需同時預測出存在交互作用且在機制RNA上存在調控關係，才能預測給定的CRM能調控特定基因。本研究開發的預測管線在auROC上的表現達到84.1%，在auPRC上的表現達到97.4%，在辨別CRM是否調控某基因方面，優於現有的計算方法和工具。這項研究不僅為理解CRM之目標基因之間的關係提供了新的見解，還為未來基因調控研究提供了有力的工具。

Transcriptional regulation in the genomes of multicellular organisms involves the interaction of transcription factors (TFs) or regulatory proteins with specific DNA sequences called cis-regulatory modules (CRMs) to control gene expression. This study aims to develop a comprehensive tool for predicting whether a CRM regulates a particular gene. Previous studies primarily determined the target genes of CRMs based on the interaction between CRMs and gene promoters. However, recent findings suggest that enhancers can target distant genes via the transcription of enhancer RNAs (eRNAs). It has also been discovered that the distance between a CRM and a gene does not directly explain their regulatory relationship. This study developed two models to predict the interactions between CRMs and genes and the regulatory relationships of CRMs on eRNAs, thereby integrating the interaction and enhancer RNA regulatory mechanisms. The study requires both interaction presence and regulatory relationships on eRNAs to predict that a given CRM can regulate a specific gene. The predictive pipeline developed in this study achieved an auROC of 84.1% and an auPRC of 97.4%, outperforming existing computational methods and tools in identifying CRM-gene regulatory relationships. This research provides new insights into CRM-target gene relationships and offers a powerful tool for future gene regulation studies.

[1]　　H. Xu, S. Zhang, X. Yi, D. Plewczynski, and M. J. Li, “Exploring 3D chromatin contacts in gene regulation: The evolution of approaches for the identification of functional enhancer-promoter interaction,” Computational and Structural Biotechnology Journal, vol. 18, pp. 558–570, 2020.
[2]　　J.E. Moore, H. E. Pratt, M. J. Purcaro, and Z. Weng, "A curated benchmark of enhancer-gene interactions for evaluating enhancer-target gene prediction methods," Genome Biol., vol. 21, no. 1, p. 17, Dec. 2020.
[3]　　K. Suryamohan and M. S. Halfon, "Identifying transcriptional cis-regulatory modules in animal genomes: Identifying transcriptional cis-regulatory modules," WIREs Dev. Biol., vol. 4, no. 2, pp. 59–84, Mar. 2015.
[4]　　J.E. Moore, H. E. Pratt, M. J. Purcaro, and Z. Weng, "A curated benchmark of enhancer-gene interactions for evaluating enhancer-target gene prediction methods," Genome Biol., vol. 21, no. 1, p. 17, Dec. 2020.
[5]　　Z. Wang, L. Zhou, S. Jiang, and W. Huang, "EPnet: A general network to predict enhancer-promoter interactions," in 2021 11th International Conference on Information Science and Technology (ICIST), Chengdu, China: IEEE, May 2021.
[6]　　Chen, Ken et al. “Capturing large genomic contexts for accurately predicting enhancer-promoter interactions.” Briefings in bioinformatics vol. 23,2 (2022)
[7]　　Andersson, Robin et al. “A unified architecture of transcriptional regulatory elements.” Trends in genetics : TIG vol. 31,8 (2015)
[8]　　T. O'Connor, C. E. Grant, M. Bodén, and T. L. Bailey, "T-Gene: improved target gene prediction," Bioinformatics, vol. 36, no. 12, pp. 3902–3904, Jun. 2020.
[9]　　A. Agarwal and L. Chen, "DeepPHiC: Predicting promoter-centered chromatin interactions using a novel deep learning approach." Bioinformatics, Volume 39, Issue 1, January 2023
[10]　　D. Özdemir and G. Gambetta, "The Role of Insulation in Patterning Gene Expression," Genes, vol. 10, no. 10, p. 767, Sep. 2019.
[11]　　S. Thomas et al., "Dynamic reprogramming of chromatin accessibility during Drosophila embryo development," Genome Biol., vol. 12, no. 5, pp. 1-17, 2011.
[12]　　T. N. Mavrich et al., "Nucleosome organization in the Drosophila genome," Nature, vol. 453, no. 7193, pp. 358-362, 2008.
[13]　　T. H. Yang, C. C. Wang, P. C. Hung, and W. S. Wu, "cis MEP: an integrated repository of genomic epigenetic profiles and cis-regulatory modules in Drosophila," BMC Syst. Biol., vol. 8, no. 4, p. S8, 2014.
[14]　　G. J. Filion et al., "Systematic Protein Location Mapping Reveals Five Principal Chromatin Types in Drosophila Cells," Cell, vol. 143, no. 2, pp. 212-224, 2010.
[15]　　W. A. Kellner, E. Ramos, K. VanBortle, N. Takenaka, and V. G. Corces, "Genome-wide phosphoacetylation of histone H3 at Drosophila enhancers and promoters," Genome Res., vol. 22, no. 6, pp. 1081-1088, 2012.
[16]　　O. Fornes et al., "JASPAR 2020: Update of the open-Access database of transcription factor binding profiles," Nucleic Acids Res., vol. 48, no. D1, pp. D87-D92, 2020.
[17]　　R. J. Schmitz, E. Grotewold, and M. Stam, "Cis-regulatory sequences in plants: Their importance, discovery, and future challenges," The Plant Cell, vol. 34, no. 2, pp. 718–741, Feb. 2022.
[18]　　D. Özdemir and G. Gambetta, "The Role of Insulation in Patterning Gene Expression," Genes, vol. 10, no. 10, p. 767, Sep. 2019.
[19]　　M. Chen, X. Liu, Q. Liu, D. Shi, and H. Li, "3D genomics and its applications in precision medicine," Cell Mol Biol Lett, vol. 28, no. 1, p. 19, Mar. 2023.
[20]　　P. J. Farnham, "Insights from genomic profiling of transcription factors," Nat. Rev. Genet., vol. 10, no. 9, pp. 605–616, Sep. 2009.
[21]　　T.-H. Yang, Y.-C. Yang, and K.-C. Tu, "regCNN: identifying Drosophila genome-wide cis-regulatory modules via integrating the local patterns in epigenetic marks and transcription factor binding motifs," Computational and Structural Biotechnology Journal, vol. 20, pp. 296–308, 2022.
[22]　　T. H. Yang, Y. H. Yu, S. H. Wu, and F. Y. Zhang, "CFA: An explainable deep learning model for annotating the transcriptional roles of cis-regulatory modules based on epigenetic codes," Comput. Biol. Med., vol. 152, p. 106375, Jan. 2023.
[23]　　B. Batut et al., "Community-driven data analysis training for biology," Cell Syst., vol. 6, no. 6, pp. 752-758, 2018.
[24]　　L. S. Gramates et al., "FlyBase: a guided tour of highlighted features," Genetics, vol. 220, no. 4, p. iyac035, 2022.
[25]　　S. Kiranyaz, O. Avci, O. Abdeljaber, T. Ince, M. Gabbouj, and D. J. Inman, "1D convolutional neural networks and applications: A survey," J. Name, vol. Volume, no. Issue, pp. Page Range, Year.
[26]　　A. Vaswani et al., "Attention is all you need," in Adv. Neural Inf. Process. Syst., pp. 5998-6008, 2017.
[27]　　V. Sartorelli and S. M. Lauberth, "Enhancer RNAs are an important regulatory layer of the epigenome," Nat. Struct. Mol. Biol., vol. 27, no. 6, pp. 521-528, Jun. 2020.
[28]　　T. Gao and J. Qian, "EnhancerAtlas 2.0: an updated resource with enhancer annotation in 586 tissue/cell types across nine species," Nucleic Acids Res., 2019.
[29]　　 R. P. Welch et al., "ChIP-Enrich: Gene set enrichment testing for ChIP-seq data," Nucleic Acids Res., vol. 42, no. 13, p. e105, 2014.
[30]　　C. T. Lee et al., "Poly-Enrich: count-based methods for gene set enrichment testing with genomic regions," NAR Genomics Bioinform., vol. 2, no. 1, p. lqz019, Mar. 2020.
[31]　　C. T. Lee, K. Wang, T. Qin, and M. A. Sartor, "Testing proximity of genomic regions to transcription start sites and enhancers complements gene set enrichment testing," Front. Genet., vol. 11, p. 199, 2020.
[32]　　T. Qin et al., "Comprehensive enhancer-target gene assignments improve gene set level interpretation of genome-wide regulatory data."Genome Biol 23, 105 (2022).
[33]　　R. E. Thurman et al., "The accessible chromatin landscape of the human genome," Nature, vol. 489, pp. 75–82, 2012.
[34] 　　Nyandwi, Jean. “The Transformer Blueprint: A Holistic Guide to the Transformer Neural Network Architecture.” Deep Learning Revision, July 2023.