Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

A unified framework for integrative study of heterogeneous gene regulatory mechanisms

Matters Arising to this article was published on 08 July 2021


Gene expression is regulated by a large variety of mechanisms. Previous studies attempting to model the quantitative relationships between gene expression levels and regulatory mechanisms have considered only one or a few mechanisms at a time, which cannot provide a full picture of the complex interactions among different mechanisms. This was partially due to the heterogeneity of the mechanisms, which involve different types of biological objects and data representations, making it hard to study them in a unified way. Here, we describe a flexible framework that can integrate very different types of data for studying their joint effects on gene expression. In this framework, domain knowledge is represented by metapaths, while the manifestations of their effects in actual data are summarized by an embedding of the biological objects in a latent space. We demonstrate the use of our framework in integrating several diverse types of data that are related to gene expression in different ways, including DNA contacts in three-dimensional genome architecture, protein–protein interactions, genomic neighbourhoods and broad chromatin accessibility domains. The modelling results reveal that these several types of data are able to model gene expression fairly well individually, but even better when integrated.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Schematic diagram of GEEK.
Fig. 2: Performance of the gene expression models.
Fig. 3: Comparing the performance of GEEK in the whole-genome and per-chromosome settings.
Fig. 4: Biological interpretations of the embeddings produced by GEEK.
Fig. 5: Performance of the gene expression models in the across-sample tests.

Data availability

Example data for testing the source code are available at The public data used for producing the results in this study and the embeddings produced from the five cell lines can be downloaded from

Code availability

Source codes are available at


  1. 1.

    Lodish, H. et al. Molecular Cell Biology 8th edn (W. H. Freeman and Company, 2016).

  2. 2.

    Boyle, E. A., Li, Y. I. & Pritchard, J. K. An expanded view of complex traits: from polygenic to omnigenic. Cell 169, 1177–1186 (2017).

    Article  Google Scholar 

  3. 3.

    Tang, J., Qu, M. & Mei, Q. PTE: predictive text embedding through large-scale heterogeneous text networks. In Proceedings of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 1165–1174 (ACM, 2015).

  4. 4.

    Sun, Y., Han, J., Yan, X., Yu, P. S. & Wu, T. PathSim: meta path-based top-k similarity search in heterogeneous information networks. Proc. VLDB Endowment 4, 992–1003 (2011).

    Article  Google Scholar 

  5. 5.

    Mikolov, T., Chen, K., Corrado, G. & Dean, J. Efficient estimation of word representations in vector space. Preprint at (2013).

  6. 6.

    Mikolov, T., Sutskever, I., Chen, K., Corrado, G. S. & Dean, J. Distributed representations of words and phrases and their compositionality. In Proceedings of Advances in Neural Information Processing Systems (NeurIPS) 3111–3119 (NIPS, 2013).

  7. 7.

    Grover, A. & Leskovec, J. node2vec: scalable feature learning for networks. In Proceedings of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 855–864 (ACM, 2016).

  8. 8.

    Zitnik, M. & Leskovec, J. Predicting multicellular function through multi-layer tissue networks. Bioinformatics 33, i190–i198 (2017).

    Article  Google Scholar 

  9. 9.

    Zeng, W., Wu, M. & Jiang, R. Prediction of enhancer–promoter interactions via natural language processing. BMC Genomics 19, 84 (2018).

    Article  Google Scholar 

  10. 10.

    Dong, Y., Chawla, N. V. & Swami, A. metapath2vec: scalable representation learning for heterogeneous networks. In Proceedings of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 135–144 (ACM, 2017).

  11. 11.

    Faruqui, M. et al. Retrofitting word vectors to semantic lexicons. In Proceedings of Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) 1606–1615 (Association for Computational Linguistics, 2015).

  12. 12.

    Vulić, I. & Mrkšić, N. Specialising word vectors for lexical entailment. In Proceedings of Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) 1134–1145 (Association for Computational Linguistics, 2018).

  13. 13.

    Xu, C. et al. Rc-net: a general framework for incorporating knowledge into word representations. In Proceedings of International Conference on Information and Knowledge Management (CIKM) 1219–1228 (ACM, 2014).

  14. 14.

    Yu, M. & Dredze, M. Improving lexical embeddings with semantic knowledge. In Annual Meeting of the Association for Computational Linguistics (ACL) (Short Papers) 545–550 (ACL, 2014).

  15. 15.

    Marcotte, E. M., Pellegrini, M., Thompson, M. J., Yeates, T. O. & Eisenberg, D. A combined algorithm for genome-wide prediction of protein function. Nature 402, 83–86 (1999).

    Article  Google Scholar 

  16. 16.

    Michalak, P. Coexpression coregulation, and cofunctionality of neighboring genes in eukaryotic genomes. Genomics 91, 243–248 (2008).

    Article  Google Scholar 

  17. 17.

    Hu, X., Shi, C. H. & Yip, K. Y. A novel method for discovering local spatial clusters of genomic regions with functional relationships from DNA contact maps. Bioinformatics 32, i111–i120 (2016).

    Article  Google Scholar 

  18. 18.

    Chen, T. & Guestrin, C. XGBoost: a scalable tree boosting system. In Proceedings of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785–794 (ACM, 2016).

  19. 19.

    Kipf, T. N. & Welling, M. Semi-supervised classification with graph convolutional networks. Preprint at (2017).

  20. 20.

    Tang, Z. et al. CTCF-mediated human 3D genome architecture reveals chromatin topology for transcription. Cell 163, 1611–1627 (2015).

    Article  Google Scholar 

  21. 21.

    Schmitt, A. D. et al. A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Rep. 17, 2042–2059 (2016).

    Article  Google Scholar 

  22. 22.

    Sima, J. et al. Identifying cis elements for spatiotemporal control of mammalian DNA replication. Cell 176, 816–830 (2019).

    Article  Google Scholar 

  23. 23.

    Ma, J. & Duan, Z. Replication timing becomes intertwined with 3D genome organization. Cell 176, 681–684 (2019).

    Article  Google Scholar 

  24. 24.

    Ernst, J. & Kellis, M. Large-scale imputation of epigenomic datasets for systematic annotation of diverse human tissues. Nat. Biotechnol. 33, 364–376 (2015).

    Article  Google Scholar 

  25. 25.

    Artetxe, M., Labaka, G., Lopez-Gazpio, I. & Agirre, E. Uncovering divergent linguistic information in word embeddings with lessons for intrinsic and extrinsic evaluation. In Proceedings of Conference on Computational Natural Language Learning (CoNLL) 282–291 (Association for Computational Linguistics, 2018).

  26. 26.

    Kiela, D., Hill, F. & Clark, S. Specializing word embeddings for similarity or relatedness. In Proceedings of Conference on Empirical Methods in Natural Language Processing (EMNLP) 2044–2048 (Association for Computational Linguistics, 2015).

  27. 27.

    Chatr-Aryamontri, A. et al. The BioGRID interaction database: 2017 update. Nucleic Acids Res. 45, D369–D379 (2017).

    Article  Google Scholar 

  28. 28.

    Harrow, J. et al. GENCODE: the reference human genome annotation for the ENCODE project. Genome Res. 22, 1760–1774 (2012).

    Article  Google Scholar 

  29. 29.

    Kundaje, A. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330 (2015).

    Article  Google Scholar 

  30. 30.

    Rao, S. S. et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 159, 1665–1680 (2014).

    Article  Google Scholar 

  31. 31.

    Barrett, T. et al. NCBI GEO: archive for functional genomics data sets—update. Nucleic Acids Res. 41, D991–D995 (2012).

    Article  Google Scholar 

  32. 32.

    Durand, N. C. et al. Juicer provides a one-click system for analyzing loop-resolution Hi-C experiments. Cell Syst. 3, 95–98 (2016).

    Article  Google Scholar 

  33. 33.

    Ay, F., Bailey, T. L. & Noble, W. S. Statistical confidence estimation for Hi-C data reveals regulatory chromatin contacts. Genome Res. 24, 999–1011 (2014).

    Article  Google Scholar 

  34. 34.

    Blondel, V. D., Guillaume, J.-L., Lambiotte, R. & Lefebvre, E. Fast unfolding of communities in large networks. Theory Exp. 2008, P10008 (2008).

    MATH  Google Scholar 

  35. 35.

    Ashburner, M. et al. Gene ontology: tool for the unification of biology. Nat. Genet. 25, 25–29 (2000).

    Article  Google Scholar 

  36. 36.

    Gene Ontology Consortium The gene ontology resource: 20 years and still GOing strong.Nucleic Acids Res 47, D330–D338 (2018).

    Article  Google Scholar 

  37. 37.

    Klopfenstein, D. et al. GOATOOLS: a Python library for gene ontology analyses. Sci. Rep. 8, 10872 (2018).

    Article  Google Scholar 

  38. 38.

    Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B Methodol. 57, 289–300 (1995).

    MathSciNet  MATH  Google Scholar 

  39. 39.

    Pedregosa, F. et al. Scikit-learn: machine learning in python. J. Mach. Learn. Res. 12, 2825–2830 (2011).

    MathSciNet  MATH  Google Scholar 

  40. 40.

    Huttlin, E. L. et al. Architecture of the human interactome defines protein communities and disease networks. Nature 545, 505–509 (2017).

    Article  Google Scholar 

  41. 41.

    Wang, Y. et al. The 3D genome browser: a web-based browser for visualizing 3D genome organization and long-range chromatin interactions. Genome Biol. 19, 151 (2018).

    Article  Google Scholar 

  42. 42.

    The ENCODE Project Consortium An integrated encyclopedia of DNA elements in the human genome.Nature 489, 57–74 (2012).

    Article  Google Scholar 

  43. 43.

    Cao, Q. et al. GEEK (Gene Expression Embedding frameworK) Demo (GM12878, Chromosome 1) (CodeOcean, 2020);

Download references


This work was supported by Hong Kong Research Grants Council General Research Funds 14170217 (C.C., D.L. and K.Y.Y.), 14203119 (A.S.L.C. and K.Y.Y.), 14200817, 15200715 and 15204116 (E.L.), Collaborative Research Funds 4045-18WF (A.S.L.C. and K.Y.Y.), 4054-16G (T.-L.L. and K.Y.Y.) and C4057-18EF (K.Y.Y.), Area of Excellence AoE/P-404/18 (E.L.), Theme-based Research Scheme T12C-714/14-R (K.Y.Y.), the Hong Kong Innovation and Technology Commission Innovative and Technology Fund ITS/310/18 (E.L.) and the Hong Kong Epigenomics Project (EpiHK). K.Y.Y. was also supported by CUHK Young Researcher Award, Outstanding Fellowship and Seed Funding for Strategic Areas.

Author information




K.Y.Y. conceived the study. Q.C. and Z.Z. designed and implemented GEEK. Q.C., Z.Z., A.X.F., Q.W. and K.Y.Y. performed data analyses. Q.C., A.X.F., T.-L.L., E.L., A.S.L.C., C.C., D.L. and K.Y.Y. interpreted the results. K.Y.Y. and Q.C. wrote the manuscript.

Corresponding author

Correspondence to Kevin Y. Yip.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary text (including algorithms 1 and 2), Table 1 and Figs. 1–23.

Supplementary Table 2

Functional enrichment analysis of gene clusters based on their embeddings. Each row represents the genes in a cluster that are annotated with the same functional term. The different columns specify the embedding setting, enrichment results and gene symbols, with their full descriptions provided in the file.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Cao, Q., Zhang, Z., Fu, A.X. et al. A unified framework for integrative study of heterogeneous gene regulatory mechanisms. Nat Mach Intell 2, 447–456 (2020).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing