Gene regulatory network inference based on causal discovery integrating with graph neural network

Abstract

Gene regulatory network (GRN) inference from gene expression data is a significant approach to understanding aspects of the biological system. Compared with generalized correlation-based methods, causality-inspired ones seem more rational to infer regulatory relationships. We propose GRINCD, a novel GRN inference framework empowered by graph representation learning and causal asymmetric learning, considering both linear and non-linear regulatory relationships. First, high-quality representation of each gene is generated using graph neural network. Then, we apply the additive noise model to predict the causal regulation of each regulator-target pair. Additionally, we design two channels and finally assemble them for robust prediction. Through comprehensive comparisons of our framework with state-of-the-art methods based on different principles on numerous datasets of diverse types and scales, the experimental results show that our framework achieves superior or comparable performance under various evaluation metrics. Our work provides a new clue for constructing GRNs, and our proposed framework GRINCD also shows potential in identifying key factors affecting cancer development.

Keywords

causal discovery / ensemble learning / gene regulatory network inference / gene regulatory networks / graph neural network / key regulators of disease development

References

[1] MarbachD, Costello JC, KuffnerR, VegaNM, PrillRJ, CamachoDM, et al. Wisdom of crowds for robust gene network inference. Nat Methods. 2012;9(8):796–804.

[2] ChaiLE, LohSK, LowST, Mohamad MS, DerisS, ZakariaZ. A review on the computational approaches for gene regulatory network construction. Comput Biol Med. 2014;48: 55–65.

[3] BanfM, RheeSY. Computational inference of gene regulatory networks: approaches, limitations and opportunities. Biochim Biophys Acta Gene Regul Mech. 2017;1860(1):41–52.

[4] MeinshausenN, Bühlmann P. Stability selection. J Roy Stat Soc B. 2010;72(4):417–73.

[5] KimD-C, LiuC, WuX, ZhangB, GaoJ. Inference of gene regulatory networks by integrating gene expressions and genetic perturbations. In: 2013 IEEE international conference on bioinformatics and biomedicine. IEEE; 2013. p. 182–7.

[6] FaithJJ, HayeteB, ThadenJT, Mogno I, WierzbowskiJ, CottarelG, et al. Large-scale mapping and validation of Escherichia coli transcriptional regulation from a compendium of expression profiles. PLoS Biol. 2007;5(1):e8.

[7] MargolinAA, Nemenman I, BassoK, WigginsC, Stolovitzky G, Dalla FaveraR, et al. Aracne: an algorithm for the reconstruction of gene regulatory networks in a mammalian cellular context. BMC Bioinf. 2006;7(Suppl 1):S7.

[8] YuJ, SmithVA, WangPP, Hartemink AJ, JarvisED. Advances to Bayesian network inference for generating causal networks from observational biological data. Bioinformatics. 2004;20(18): 3594–603.

[9] BalovN, Salzman P. Catnet: categorical Bayesian network inference. 2022. Available from the website of rdrr.io.

[10] TsamardinosI, Aliferis CF, StatnikovA. Time and sample efficient discovery of Markov blankets and direct causal relations. In: Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining; 2003. p. 673–8.

[11] PearlJ. Bayesian Networks. In: Alhajj, R, Rokne, J, editors. Encyclopedia of social network analysis and mining. New York, NY: Springer; 2011.

[12] QiuX, Rahimzamani A, WangL, RenB, MaoQ, DurhamT, et al. Inferring causal gene regulatory networks from coupled single-cell expression dynamics using scribe. Cell Syst. 2020;10(3):265–74.e211.

[13] Huynh-ThuVA, Irrthum A, WehenkelL, GeurtsP. Inferring regulatory networks from expression data using tree-based methods. PLoS One. 2010;5(9):e12776.

[14] HauryA-C, Mordelet F, Vera-LiconaP, VertJ-P. Tigress: trustful inference of gene regulation using stability selection. BMC Syst Biol. 2012;6:1–17.

[15] FanY, MaX. Gene regulatory network inference using 3D convolutional neural network. Proc AAAI Conf Artif Intell. 2021;35(1):99–106.

[16] WangJ, MaA, MaQ, XuD, JoshiT. Inductive inference of gene regulatory network using supervised and semi-supervised graph neural networks. Comput Struct Biotechnol J. 2020;18: 3335–43.

[17] GrangerCW. Investigating causal relations by econometric models and cross-spectral methods. Econometrica: J Econom Soc. 1969;37(3):424–38.

[18] SugiharaG, MayR, YeH, HsiehC-h, DeyleE, Fogarty M, et al. Detecting causality in complex ecosystems. Science. 2012; 338(6106):496–500.

[19] SpirtesP, Glymour CN, ScheinesR, HeckermanD. Causation, prediction, and search. MIT press; 2000.

[20] BelyaevaA, Squires C, UhlerC. Dci: learning causal differences between gene regulatory networks. Bioinformatics. 2021;37(18): 3067–9.

[21] LeTD, LiuL, TsykinA, Goodall GJ, LiuB, SunBY, et al. Inferring microrna-mrna causal regulatory relationships from expression data. Bioinformatics. 2013;29(6):765–71.

[22] ZhengX, AragamB, RavikumarPK, Xing EP. Dags with no tears: continuous optimization for structure learning. Adv Neural Inf Process Syst. 2018;31.

[23] ShenX, MaS, VemuriP, Simon G, Alzheimer’s Disease Neuroimaging I, AisenP, et al. Challenges and opportunities with causal discovery algorithms: application to alzheimer’s pathophysiology. Sci Rep. 2020;10(1):2975.

[24] HoyerP, Janzing D, MooijJM, PetersJ, Schölkopf B. Nonlinear causal discovery with additive noise models. Adv Neural Inf Process Syst. 2008;21.

[25] ZhangK, Hyvarinen A. On the identifiability of the postnonlinear causal model; 2012. arXiv preprint arXiv:12052599.

[26] ShinH-C, RothHR, GaoM, LuL, XuZ, NoguesI, et al. Deep convolutional neural networks for computer-aided detection: cnn architectures, dataset characteristics and transfer learning. IEEE Trans Med Imag. 2016;35(5):1285–98.

[27] DefferrardM, Bresson X, VandergheynstP. Convolutional neural networks on graphs with fast localized spectral filtering. Adv Neural Inf Process Syst. 2016;29.

[28] HamiltonW, YingZ, LeskovecJ. Inductive representation learning on large graphs. Adv Neural Inf Process Syst. 2017;30.

[29] WangY, LiuJ. A sparse fireworks algorithm for gene regulatory network reconstruction based on fuzzy cognitive maps. In: 2019 IEEE congress on evolutionary computation (CEC). IEEE; 2019. p. 1188–94

[30] VeličkovićP, CucurullG, Casanova A, RomeroA, LioP, BengioY. Graph attention networks; 2017. arXiv preprint arXiv:171010903.

[31] JanzingD, Steudel B, ShajarisalesN, SchölkopfB. Justifying information-geometric causal inference. Measur Complex. 2015:253–65.

[32] BlöbaumP, Janzing D, WashioT, ShimizuS, Schölkopf B. Cause-effect inference by comparing regression errors. In: International conference on artificial intelligence and statistics. PMLR; 2018. p. 900–9.

[33] GuoYP, ShaoL, WangL, Chen MY, ZhangW, HuangWH. Bioconversion variation of ginsenoside ck mediated by human gut microbiota from healthy volunteers and colorectal cancer patients. Chin Med. 2021;16(1):28.

[34] GyorffyB, MolnarB, LageH, Szallasi Z, EklundAC. Evaluation of microarray preprocessing algorithms based on concordance with RT-PCR in clinical samples. PLoS One. 2009;4(5):e5645.

[35] GalambO, Gyorffy B, SiposF, SpisakS, NemethAM, MihellerP, et al. Inflammation, adenoma and cancer: objective classification of colon biopsy specimens with gene expression signature. Dis Markers. 2008;25:1–16.

[36] GalambO, Wichmann B, SiposF, SpisakS, Krenacs T, TothK, et al. Dysplasia-carcinoma transition specific transcripts in colonic biopsy samples. PLoS One. 2012;7(11):e48547.

[37] KimP, ParkA, HanG, SunH, JiaP, ZhaoZ. Tissgdb: tissuespecific gene database in cancer. Nucleic Acids Res. 2018; 46(D1):D1031–8.

[38] HanH, ChoJ-W, LeeS, YunA, KimH, BaeD, et al. Trrust v2: an expanded reference database of human and mouse transcriptional regulatory interactions. Nucleic Acids Res. 2018; 46(D1):D380–6.

[39] Liu,X, Ge,J, Tian,L, Zhang, J, Li,H, Ding,Y, et al. Identification of wwtr1 as an immune infiltration-correlated prognostic biomarker in colon cancer. 2022. Available from the website of semanticscholar.

[40] Wang,W-D, Wang,L-C, Liu,M-X, Lu, Y-J, and Jiang, B. Transcription factors e2f2/3/4 as possible colorectal cancer prognostic biomarkers. 2021. Available from the website of semanticscholar.

[41] WangL, Zhang M-X, ZhangM-F, TuZ-W. Zbtb7a functioned as an oncogene in colorectal cancer. BMC Gastroenterol. 2020; 20:1–7.

[42] SafranM, RosenN, TwikM, BarShir R, SteinTI, DaharyD, et al. The genecards suite. In: Practical guide to life science databases. Springer; 2021. p. 27–56.

[43] LipinskiKA, Britschgi C, SchraderK, ChristinatY, Frischknecht L, KrekW. Colorectal cancer cells display chaperone dependency for the unconventional prefoldin uri1. Oncotarget. 2016;7(20):29635–47.

[44] SchaffterT, Marbach D, FloreanoD. Genenetweaver: in silico benchmark generation and performance profiling of network inference methods. Bioinformatics. 2011;27(16):2263–70.

[45] ÄijöT, Lähdesmäki H. Learning gene regulatory networks from gene expression measurements using non-parametric molecular kinetics. Bioinformatics. 2009;25(22):2937–44.

[46] BonneauR, ReissDJ, ShannonP, Facciotti M, HoodL, BaligaNS, et al. The inferelator: an algorithm for learning parsimonious regulatory networks from systems-biology data sets de novo. Genome Biol. 2006;7(5):R36.

[47] Gama-CastroS, Salgado H, Peralta-GilM, Santos-ZavaletaA, Muniz-Rascado L, Solano-LiraH, et al. Regulondb version 7.0: transcriptional regulation of Escherichia coli k-12 integrated within genetic sensory response units (gensor units). Nucleic Acids Res. 2010;39(Database):D98–105.

[48] HarbisonCT, GordonDB, LeeTI, Rinaldi NJ, MacisaacKD, DanfordTW, et al. Transcriptional regulatory code of a eukaryotic genome. Nature. 2004;431(7004):99–104.

[49] MacIsaacKD, WangT, GordonDB, Gifford DK, StormoGD, FraenkelE. An improved map of conserved regulatory sites for saccharomyces cerevisiae. BMC Bioinf. 2006;7(1):113.

[50] PratapaA, Jalihal AP, LawJN, BharadwajA, MuraliTM. Benchmarking algorithms for gene regulatory network inference from single-cell transcriptomic data. Nat Methods. 2020;17(2):147–54.

[51] ZouZ, OhtaT, MiuraF, Oki S. Chip-atlas 2021 update: a data-mining suite for exploring epigenomic landscapes by fully integrating chip-seq, atac-seq and bisulfite-seq data. Nucleic Acids Res. 2022;50(W1):W175–82.

[52] SzklarczykD, GableAL, LyonD, Junge A, WyderS, Huerta-CepasJ, et al. String v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–13.

[53] KingmaDP, BaJ. Adam: a method for stochastic optimization; 2014. arXiv preprint arXiv:14126980.

[54] ShimizuS. Lingam: non-Gaussian methods for estimating causal structures. Behaviormetrika. 2014;41(1):65–98.

[55] SchulzE, Speekenbrink M, KrauseA. A tutorial on Gaussian process regression: modelling, exploring, and exploiting functions. J Math Psychol. 2018;85:1–16.

[56] GrettonA, Fukumizu K, TeoC, SongL, Schölkopf B, SmolaA. A kernel statistical test of independence. Adv Neural Inf Process Syst. 2007;20.

[57] EmersonP. The original borda count and partial voting. Soc Choice Welfare. 2013;40(2):353–8.

[58] StolovitzkyG, PrillRJ, CalifanoA. Lessons from the dream2 challenges. Ann N Y Acad Sci. 2009;1158(1):159–95.

[59] YeungKY, Bumgarner RE, RafteryAE. Bayesian model averaging: development of an improved multi-class, gene selection and classification tool for microarray data. Bioinformatics. 2005;21(10):2394–402.

RIGHTS & PERMISSIONS

2023 The Authors. Quantitative Biology published by John Wiley & Sons Australia, Ltd on behalf of Higher Education Press.