Letter | Published:

Detection and replication of epistasis influencing transcription in humans

Nature volume 508, pages 249253 (10 April 2014) | Download Citation


Epistasis is the phenomenon whereby one polymorphism’s effect on a trait depends on other polymorphisms present in the genome. The extent to which epistasis influences complex traits1 and contributes to their variation2,3 is a fundamental question in evolution and human genetics. Although often demonstrated in artificial gene manipulation studies in model organisms4,5, and some examples have been reported in other species6, few examples exist for epistasis among natural polymorphisms in human traits7,8. Its absence from empirical findings may simply be due to low incidence in the genetic control of complex traits2,3, but an alternative view is that it has previously been too technically challenging to detect owing to statistical and computational issues9. Here we show, using advanced computation10 and a gene expression study design, that many instances of epistasis are found between common single nucleotide polymorphisms (SNPs). In a cohort of 846 individuals with 7,339 gene expression levels measured in peripheral blood, we found 501 significant pairwise interactions between common SNPs influencing the expression of 238 genes (P < 2.91 × 10−16). Replication of these interactions in two independent data sets11,12 showed both concordance of direction of epistatic effects (P = 5.56 × 10−31) and enrichment of interaction P values, with 30 being significant at a conservative threshold of P < 9.98 × 10−5. Forty-four of the genetic interactions are located within 5 megabases of regions of known physical chromosome interactions13 (P = 1.8 × 10−10). Epistatic networks of three SNPs or more influence the expression levels of 129 genes, whereby one cis-acting SNP is modulated by several trans-acting SNPs. For example, MBNL1 is influenced by an additive effect at rs13069559, which itself is masked by trans-SNPs on 14 different chromosomes, with nearly identical genotype–phenotype maps for each cistrans interaction. This study presents the first evidence, to our knowledge, for many instances of segregating common polymorphisms interacting to influence human traits.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.


Gene Expression Omnibus

Data deposits

Gene expression data is available at the Gene Expression Omnibus under accession code GSE53195.


  1. 1.

    & Epistasis: too often neglected in complex trait studies? Nature Rev. Genetics 5, 618–625 (2004)

  2. 2.

    , & Data and theory point to mainly additive genetic variance for complex traits. PLoS Genet. 4, e1000008 (2008)

  3. 3.

    On epistasis: why it is unimportant in polygenic directional selection. Phil. Trans. R. Soc. B 365, 1241–1244 (2010)

  4. 4.

    et al. The genetic landscape of a cell. Science 327, 425–431 (2010)

  5. 5.

    , , , & Finding the sources of missing heritability in a yeast cross. Nature 234–237 (2013)

  6. 6.

    , , , & Epistasis and the release of genetic variation during long-term selection. Nature Genetics 38, 418–420 (2006)

  7. 7.

    et al. A genome-wide association study identifies new psoriasis susceptibility loci and an interaction between HLA-C and ERAP1. Nature Genetics 42, 985–990 (2010)

  8. 8.

    et al. Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nature Genetics 43, (2011)

  9. 9.

    Detecting gene–gene interactions that underlie human diseases. Nature Rev. Genetics 10, 392–404 (2009)

  10. 10.

    , , & EpiGPU: exhaustive pairwise epistasis scans parallelized on consumer level graphics cards. Bioinformatics 27, 1462–1465 (2011)

  11. 11.

    The Estonian Genome Project. Drug Dev. Res. 62, 97–101 (2004)

  12. 12.

    et al. Trans-eQTLs reveal that independent genetic variants associated with a complex phenotype converge on intermediate genes, with a major role for the HLA. PLoS Genetics 7, e1002197 (2011)

  13. 13.

    et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326, 289–293 (2009)

  14. 14.

    , , & Five years of GWAS discovery. Am. J. Hum. Genet. 90, 7–24 (2012)

  15. 15.

    , , , & Darwinian evolution can follow only very few mutational paths to fitter proteins. Science 312, 111–114 (2006)

  16. 16.

    , , , & a Epistasis as the primary factor in molecular evolution. Nature 490, 535–538 (2012)

  17. 17.

    Linkage disequilibrium and association mapping. Annu. Rev. Genomics Hum. Genet. 9, 129–142 (2008)

  18. 18.

    , & An evolutionary perspective on epistasis and the missing heritability. PLoS Genet. 9, e1003295 (2013)

  19. 19.

    , & Genome-wide strategies for detecting multiple loci that influence complex diseases. Nature Genet. 37, 413–417 (2005)

  20. 20.

    et al. Hundreds of variants clustered in genomic loci and biological pathways affect human height. Nature 467, 832–838 (2010)

  21. 21.

    et al. Genetics of gene expression surveyed in maize, mouse and man. Nature 422, 297–302 (2003)

  22. 22.

    et al. Congruence of additive and non-additive effects on gene expression estimated from pedigree and SNP data. PLoS Genet. 9, e1003502 (2013)

  23. 23.

    et al. The Brisbane Systems Genetics Study: genetical genomics meets complex trait genetics. PLoS ONE 7, e35430 (2012)

  24. 24.

    et al. Blood-informative transcripts define nine common axes of peripheral blood gene expression. PLoS Genet. 9, e1003362 (2013)

  25. 25.

    An extension of the concept of partitioning hereditary variance for analysis of covariances among relatives when epistasis is present. Genetics 39, 859–882 (1954)

  26. 26.

    et al. Muscleblind proteins regulate alternative splicing. EMBO J. 23, 3103–3112 (2004)

  27. 27.

    et al. Chromatin marks identify critical cell types for fine mapping complex trait variants. Nature Genet. 45, 124–130 (2013)

  28. 28.

    , , & Unsupervised pattern discovery in human chromatin structure through genomic segmentation. Nature Methods 9, 473–476 (2012)

  29. 29.

    et al. Integration of Hi-C and ChIP-seq data reveals distinct types of chromatin linkages. Nucleic Acids Res. 40, 7690–7704 (2012)

  30. 30.

    , & Transcription factories. Front. Genet. 3, 221 (2012)

  31. 31.

    et al. Common variants in the trichohyalin gene are associated with straight hair in Europeans. Am. J Hum. Genet. 85, 750–755 (2009)

  32. 32.

    , , & GenABEL: an R library for genome-wide association analysis. Bioinformatics 23, 1294–1296 (2007)

  33. 33.

    et al. Genome partitioning of genetic variation for complex traits using common SNPs. Nature Genet. 43, 519–525 (2011)

  34. 34.

    et al. MixupMapper: correcting sample mix-ups in genome-wide datasets increases power to detect small genetic effects. Bioinformatics 27, 2104–2111 (2011)

  35. 35.

    Improved likelihood ratio tests for complete contingency tables. Biometrika 63, 33–37 (1976)

  36. 36.

    , & How to perform meaningful estimates of genetic effects. PLoS Genet. 4, e1000062 (2008)

  37. 37.

    et al. GWAS of 126,559 individuals identifies genetic variants associated with educational attainment. Science 340, 1467–1471 (2013)

  38. 38.

    DNA binding sites: representation and discovery. Bioinformatics 16, 16–23 (2000)

  39. 39.

    et al. oPOSSUM: identification of over-represented transcription factor binding sites in co-expressed genes. Nucleic Acids Res. 33, 3154–3164 (2005)

Download references


We are grateful to the volunteers for their participation in these studies. We thank B. Hill, C. Haley and L. Ronnegard for discussions and comments. This work could not have been completed without access to high performance GPGPU compute clusters. We acknowledge iVEC for the use of advanced computing resources located at iVEC@UWA (http://www.ivec.org), and the Multi-modal Australian Sciences Imaging and Visualisation Environment (MASSIVE) (http://www.massive.org.au). We also thank J. Carroll and I. Porebski from the Queensland Brain Institute Information Technology Group for HPC support. The University of Queensland group is supported by the Australian National Health and Medical Research Council (NHMRC) grants 389892, 496667, 613601, 1010374 and 1046880, the Australian Research Council (ARC) grant (DE130100691), and by National Institutes of Health (NIH) grants GM057091 and GM099568. The QIMR researchers acknowledge funding from the Australian National Health and Medical Research Council (grants 241944, 389875, 389891, 389892, 389938, 442915, 442981, 496739, 496688 and 552485), and the National Institutes of Health (grants AA07535, AA10248, AA014041, AA13320, AA13321, AA13326 and DA12854). We thank A. Caracella and L. Bowdler for technical assistance with the micro-array hybridisations. The CHDWB study funding support from the Georgia Institute of Technology Research Foundation. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. The Fehrmann study was supported by grants from the Celiac Disease Consortium (an innovative cluster approved by the Netherlands Genomics Initiative and partly funded by the Dutch Government (grant BSIK03009)), the Netherlands Organization for Scientific Research (NWO-VICI grant 918.66.620, NWO-VENI grant 916.10.135 to L.F.), the Dutch Digestive Disease Foundation (MLDS WO11-30), and a Horizon Breakthrough grant from the Netherlands Genomics Initiative (grant 92519031 to L.F.). This project was supported by the Prinses Beatrix Fonds, VSB fonds, H. Kersten and M. Kersten (Kersten Foundation), The Netherlands ALS Foundation, and J.R. van Dijk and the Adessium Foundation. The research leading to these results has received funding from the European Communitys Health Seventh Framework Programme (FP7/2007-2013) under grant agreement 259867. The EGCUT study received targeted financing from the Estonian Government SF0180142s08, Center of Excellence in Genomics (EXCEGEN) and University of Tartu (SP1GVARENG). We acknowledge EGCUT technical personnel, especially V. Soo and S. Smit. Data analyses were carried out in part in the High Performance Computing Center of University of Tartu.

Author information

Author notes

    • Grant W. Montgomery
    • , Peter M. Visscher
    •  & Joseph E. Powell

    These authors contributed equally to this work.


  1. Queensland Brain Institute, University of Queensland, Brisbane, Queensland 4072, Australia

    • Gibran Hemani
    • , Konstantin Shakhbazov
    • , Allan F. McRae
    • , Jian Yang
    • , Peter M. Visscher
    •  & Joseph E. Powell
  2. University of Queensland Diamantina Institute, University of Queensland, Princess Alexandra Hospital, Brisbane, Queensland 4072, Australia

    • Gibran Hemani
    • , Konstantin Shakhbazov
    • , Allan F. McRae
    • , Peter M. Visscher
    •  & Joseph E. Powell
  3. Department of Genetics, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9700 RB Groningen, The Netherlands

    • Harm-Jan Westra
    •  & Lude Franke
  4. Estonian Genome Center, University of Tartu, Tartu 51010, Estonia

    • Tonu Esko
    •  & Andres Metspalu
  5. Medical and Population Genetics, Broad Institute, Cambridge, Massachusetts 02142, USA

    • Tonu Esko
  6. Divisions of Endocrinology, Children's Hospital, Boston, Massachusetts 02115, USA

    • Tonu Esko
  7. Queensland Institute of Medical Research, Brisbane, Queensland 4006, Australia

    • Anjali K. Henders
    • , Nicholas G. Martin
    •  & Grant W. Montgomery
  8. School of Biology and Centre for Integrative Genomics, Georgia Institute of Technology, Atlanta, Georgia 30332, USA

    • Greg Gibson


  1. Search for Gibran Hemani in:

  2. Search for Konstantin Shakhbazov in:

  3. Search for Harm-Jan Westra in:

  4. Search for Tonu Esko in:

  5. Search for Anjali K. Henders in:

  6. Search for Allan F. McRae in:

  7. Search for Jian Yang in:

  8. Search for Greg Gibson in:

  9. Search for Nicholas G. Martin in:

  10. Search for Andres Metspalu in:

  11. Search for Lude Franke in:

  12. Search for Grant W. Montgomery in:

  13. Search for Peter M. Visscher in:

  14. Search for Joseph E. Powell in:


G.H., J.E.P., P.M.V. and G.W.M. conceived and designed the study. G.H., J.E.P., K.S., H.-J.W. and J.Y. performed the analysis. T.E. and A.M. provided the EGCUT data. A.K.H., A.F.M., G.W.M., N.G.M. and J.E.P. provided the BSGS data. G.G. provided the CHDWB data. H.-J.W. and L.F. provided the Fehrmann data. G.H. and J.E.P. wrote the manuscript with the participation of all authors.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Gibran Hemani.

Supplementary information

PDF files

  1. 1.

    Supplementary Information

    This file contains Supplementary Figures 1-17 and Supplementary Tables 1-5.

About this article

Publication history






Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.