Journal of Zhejiang University

Journal of Zhejiang University SCIENCE B 2026 Vol.27 No.10 P.1129-1139

Integrative graph deep learning with directional decoding for robust gene regulatory network inference from single-cell transcriptomics

Author(s): Binhua TANG,Yingying FENG,Xinyu GAO,Mengyao MAO,Yujia ZHANG
Affiliation(s): 1. Key Laboratory of Maritime Intelligent Cyberspace Technology (Ministry of Education of China), Hohai University, Nanjing 213200, China more
Corresponding email(s): bh.tang@hhu.edu.cn
Key Words: Single-cell RNA sequencing (scRNA-seq), Gene regulatory network (GRN) inference, Local-aware graph aggregation (LGA), Node reconstruction autoencoder, Asymmetric relation decoder

<<< Previous Article \|Next Article >>>

Binhua TANG, Yingying FENG, Xinyu GAO, Mengyao MAO, Yujia ZHANG. Integrative graph deep learning with directional decoding for robust gene regulatory network inference from single-cell transcriptomics[J]. Journal of Zhejiang University Science B, 2026, 27(10): 1129-1139.

@article{title="Integrative graph deep learning with directional decoding for robust gene regulatory network inference from single-cell transcriptomics",
author="Binhua TANG, Yingying FENG, Xinyu GAO, Mengyao MAO, Yujia ZHANG",
journal="Journal of Zhejiang University Science B",
volume="27",
number="10",
pages="1129-1139",
year="2026",
publisher="Zhejiang University Press & Springer",
doi="10.1631/jzus.B2500424"
}

%0 Journal Article
%T Integrative graph deep learning with directional decoding for robust gene regulatory network inference from single-cell transcriptomics
%A Binhua TANG
%A Yingying FENG
%A Xinyu GAO
%A Mengyao MAO
%A Yujia ZHANG
%J Journal of Zhejiang University SCIENCE B
%V 27
%N 10
%P 1129-1139
%@ 1673-1581
%D 2026
%I Zhejiang University Press & Springer
%DOI 10.1631/jzus.B2500424

TY - JOUR
T1 - Integrative graph deep learning with directional decoding for robust gene regulatory network inference from single-cell transcriptomics
A1 - Binhua TANG
A1 - Yingying FENG
A1 - Xinyu GAO
A1 - Mengyao MAO
A1 - Yujia ZHANG
J0 - Journal of Zhejiang University Science B
VL - 27
IS - 10
SP - 1129
EP - 1139
%@ 1673-1581
Y1 - 2026
PB - Zhejiang University Press & Springer
ER -
DOI - 10.1631/jzus.B2500424

Abstract
Chinese Summary
Academic Network
Reviewer Comment

Abstract: While gene regulatory networks (GRNs) are fundamental to understanding complex cellular mechanisms, the accurate inference from single-cell RNA-sequencing (scRNA-seq) data is hindered by the inherent noise, transcriptional dropout events, and dynamic nature of gene regulation. Current computational approaches often overlook long-range regulatory interactions, lack explicit modeling of regulatory directionality, and suffer from limited generalizability. To address these limitations, we introduce scGRAIL, a supervised deep learning framework that combines inductive local-aware graph aggregation (LGA) with an asymmetric relation decoder to enable robust GRN inference. Within this framework, LGA performs multi-hop neighborhood aggregation on a transcription factor (TF)–target graph, jointly leveraging topological structure and gene expression to capture long-range regulatory dependencies. A node reconstruction autoencoder (NAE) imposes gene expression reconstruction constraints on the latent embeddings, thereby enhancing robustness to noise and data heterogeneity and improving generalization. An asymmetric difference mechanism is introduced during decoding for distinct modeling of TFs and target genes; by computing an ordered difference vector between their embeddings, the model explicitly captures the directionality and target specificity of TF–target regulatory interactions, facilitating more accurate identification of potential causal regulatory relationships. Comprehensive evaluations demonstrate the model’s superiority: it achieves optimal performance on 86.4% of benchmark datasets (38/44), surpassing state-of-the-art methods. Notably, scGRAIL achieves a 4.25% increase in average area under the receiver operating characteristic curve (AUROC) across four real-world networks compared to the graph contrastive link prediction (GCLink) model, yielding a 17.64% AUROC improvement for the TF500+ Non-Specific network in the mouse hematopoietic stem cells with lymphoid-lineage (mHSC-L) dataset, demonstrating its resilience to noise and scalability for large-scale GRNs. These results establish scGRAIL as a transformative tool for uncovering gene regulatory logic in single-cell transcriptomics. Overall, scGRAIL enhances structural modeling and directional discrimination in single-cell GRN inference by using multi-hop graph aggregation to capture long-range regulatory dependencies, expression reconstruction constraints to improve cross-dataset robustness, and an asymmetric decoding mechanism to explicitly model the regulatory directionality between TFs and target genes.

融合方向性解码的图深度学习框架：面向单细胞转录组的鲁棒基因调控网络推断

汤斌华^1,2, 冯盈盈¹, 高欣宇¹, 毛梦瑶¹, 张玉嘉¹
¹河海大学，教育部海上智能网信技术重点实验室，中国南京，213200
²复旦大学，上海市数据科学重点实验室，中国上海，200438
摘要：基因调控网络（GRNs）是揭示细胞内复杂调控机制的重要工具，但基于单细胞RNA测序（scRNA-seq）数据的GRN推断常受噪声、转录本"dropout"以及调控关系动态性等因素影响而难以准确建模。现有方法往往忽略长程调控依赖，缺乏对调控方向性的显式刻画，并存在泛化能力不足等问题。为此，本文提出一种有监督的深度学习框架scGRAIL，将归纳式局部感知图聚合（LGA）与非对称关系解码器相结合，实现鲁棒的GRN推断。在该框架中，LGA在转录因子-靶标（TF-target）图上进行多跳邻域聚合，联合利用网络拓扑结构与基因表达信息以捕获长程调控依赖；同时引入节点重构自编码器（NAE）对潜在表示施加表达重构约束，从而增强对噪声和数据异质性的鲁棒性，并提升跨数据集泛化能力。解码阶段设计非对称差分机制，通过计算TF与Target嵌入的有序差分向量，对TF-target调控关系的方向性与靶基因特异性进行显式建模，从而更准确识别潜在的因果调控关系。综合实验结果表明，scGRAIL在44个基准数据集中的38个（86.4%）取得最优性能，整体优于现有先进方法；在四个真实网络上，相较GCLink的平均受试者工作特征曲线下面积（AUROC）提升4.25%，其中在Non-Specific网络，mHSC-L中TF500+数据集上AUROC提升17.64%，体现出对噪声的强鲁棒性及大规模网络的可扩展性。综上，scGRAIL通过多跳图聚合捕获长程依赖，以表达重构增强跨数据集鲁棒性，并利用非对称方向性解码提升调控判别能力，为单细胞转录组中的基因调控逻辑解析提供了有效工具。

关键词：单细胞RNA测序（scRNA-seq）；基因调控网络（GRN）推断；局部感知图聚合（LGA）；节点重构自编码器；非对称关系解码器

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference

[1]Chen GY, Liu ZP, 2022. Graph attention network for link prediction of gene regulations from single-cell RNA-sequencing data. Bioinformatics, 38(19):4522-4529.

[2]Cuevas-Diaz Duran R, Wei HC, Wu JQ, 2024. Data normalization for addressing the challenges in the analysis of single-cell transcriptomic datasets. BMC Genomics, 25:444.

[3]Gan YL, Yu JC, Xu GW, et al., 2024. Inferring gene regulatory networks from single-cell transcriptomics based on graph embedding. Bioinformatics, 40(5):btae291.

[4]Hamilton WL, Ying R, Leskovec J, 2017. Inductive representation learning on large graphs. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA, p.1025-1035.

[5]Huynh-Thu VA, Irrthum A, Wehenkel L, et al., 2010. Inferring regulatory networks from expression data using tree-based methods. PLoS One, 5(9):e12776.

[6]KC K, Li R, Cui F, et al., 2019. GNE: a deep learning framework for gene network inference by aggregating biological information. BMC Syst Biol, 13(Suppl 2):38.

[7]Liu W, Teng ZJ, Li ZJ, et al., 2024. CVGAE: a self-supervised generative method for gene regulatory network inference using single-cell RNA sequencing data. Interdiscip Sci, 16(4):990-1004.

[8]Mao G, Pang ZB, Zuo K, et al., 2023. Predicting gene regulatory links from single-cell RNA-seq data using graph neural networks. Brief Bioinform, 24(6):bbad414.

[9]Matsumoto H, Kiryu H, Furusawa C, et al., 2017. SCODE: an efficient regulatory network inference algorithm from single-cell RNA-seq during differentiation. Bioinformatics, 33(15):2314-2321.

[10]Moerman T, Aibar Santos S, Bravo González-Blas C, et al., 2019. GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks. Bioinformatics, 35(12):2159-2161.

[11]Papili Gao N, Ud-Dean SMM, Gandrillon O, et al., 2018. SINCERITIES: inferring gene regulatory networks from time-stamped single cell transcriptional expression profiles. Bioinformatics, 34(2):258-266.

[12]Shu HT, Zhou JT, Lian QY, et al., 2021. Modeling gene regulatory networks using neural network architectures. Nat Comput Sci, 1(7):491-501.

[13]Wang JC, Chen YJ, Zou Q, 2023. Inferring gene regulatory network from single-cell transcriptomes with graph autoencoder model. PLoS Genet, 19(9):e1010942.

[14]Wei PJ, Jin HW, Gao Z, et al., 2025. GAEDGRN: reconstruction of gene regulatory networks based on gravity-inspired graph autoencoders. Brief Bioinform, 26(3):bbaf232.

[15]Xu J, Zhang AD, Liu F, et al., 2023. STGRNS: an interpretable transformer-based method for inferring gene regulatory networks from single-cell transcriptomic data. Bioinformatics, 39(4):btad165.

[16]Yu WM, Lin ZR, Lan MF, et al., 2025. GCLink: a graph contrastive link prediction framework for gene regulatory network inference. Bioinformatics, 41(3):btaf074.

[17]Yuan L, Zhao L, Jiang YF, et al., 2024. scMGATGRN: a multiview graph attention network–based method for inferring gene regulatory networks from single-cell transcriptomic data. Brief Bioinform, 25(6):bbae526.

[18]Yuan Y, Bar-Joseph Z, 2019. Deep learning for inferring gene relationships from single-cell expression data. Proc Natl Acad Sci USA, 116(52):27151-27158.

[19]Zhao MY, He WY, Tang JJ, et al., 2022. A hybrid deep learning framework for gene regulatory network inference from single-cell transcriptomic data. Brief Bioinform, 23(2):bbab568.

[20]Zheng RQ, Li M, Chen X, et al., 2019. BiXGBoost: a scalable, flexible boosting-based method for reconstructing gene regulatory networks. Bioinformatics, 35(11):1893-1900.

Full Text: <1594>