The architecture of gene regulatory variation across multiple human tissues: the MuTHER study.
Nica AC, Parts L, Glass D, Nisbet J, Barrett A, Sekowska M, Travers M, Potter S, Grundberg E, Small K, Hedman AK, Bataille V, Tzenova Bell J, Surdulescu G, Dimas AS, Ingle C, Nestle FO, di Meglio P, Min JL, Wilk A, Hammond CJ, Hassanali N, Yang TP, Montgomery SB, O'Rahilly S, Lindgren CM, Zondervan KT, Soranzo N, Barroso I, Durbin R, Ahmadi K, Deloukas P, McCarthy MI, Dermitzakis ET, Spector TD, MuTHER Consortium. · 2011
Read The architecture of gene regulatory variation across multiple human tissues: the MuTHER study. by Nica AC, Parts L, Glass D, Nisbet J, Barrett A, Sekowska M, Travers M, Potter S, Grundberg E, Small K, Hedman AK, Bataille V, Tzenova Bell J, Surdulescu G, Dimas AS, Ingle C, Nestle FO, di Meglio P, Min JL, Wilk A, Hammond CJ, Hassanali N, Yang TP, Montgomery SB, O'Rahilly S, Lindgren CM, Zondervan KT, Soranzo N, Barroso I, Durbin R, Ahmadi K, Deloukas P, McCarthy MI, Dermitzakis ET, Spector TD, MuTHER Consortium. (2011) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc3033383.txt.
Introduction
Nica and the MuTHER consortium explored in 2011 how gene control differs across human tissues. They studied healthy female twins and took fat, skin, and blood cell samples from the same women at the same time. The question was whether DNA variants that tune gene activity act the same everywhere, or switch roles from tissue to tissue.
Thousands of tuning variants were found in each tissue. A statistical cleanup that removes hidden noise more than doubled the discoveries, and twin checks confirmed nearly all of them. About three in ten variants acted across all three tissues, while a similar share acted in only one. Even shared variants often changed strength between tissues. The authors argue that judging tissue specificity needs effect sizes, not just significance cutoffs.
Catalog record
| CREATOR | Nica AC, Parts L, Glass D, Nisbet J, Barrett A, Sekowska M, Travers M, Potter S, Grundberg E, Small K, Hedman AK, Bataille V, Tzenova Bell J, Surdulescu G, Dimas AS, Ingle C, Nestle FO, di Meglio P, Min JL, Wilk A, Hammond CJ, Hassanali N, Yang TP, Montgomery SB, O'Rahilly S, Lindgren CM, Zondervan KT, Soranzo N, Barroso I, Durbin R, Ahmadi K, Deloukas P, McCarthy MI, Dermitzakis ET, Spector TD, MuTHER Consortium. |
|---|---|
| YEAR | 2011 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Architecture of Gene Regulatory Variation across Multiple Human Tissues: The MuTHER Study By Alexandra C Nica, Leopold Parts, Daniel Glass, James Nisbet, Amy Barrett, Magdalena Sekowska, Mary Travers, Simon Potter, Elin Grundberg, Kerrin Small, Åsa K Hedman, Veronique Bataille Published in: PLoS Genetics DOI: 10.1371/journal.pgen.1002003 ABSTRACT Abstract While there have been studies exploring regulatory variation in one or more tissues, the complexity of tissue-specificity in multiple primary tissues is not yet well understood. We explore in depth the role of cis-regulatory variation in three human tissues: lymphoblastoid cell lines (LCL), skin, and fat. The samples (156 LCL, 160 skin, 166 fat) were derived simultaneously from a subset of well-phenotyped healthy female twins of the MuTHER resource. We discover an abundance of cis-eQTLs in each tissue similar to previous estimates (858 or 4.7% of genes). In addition, we apply factor analysis (FA) to remove effects of latent variables, thus more than doubling the number of our discoveries (1,822 eQTL genes). The unique study design (Matched Co-Twin Analysis—MCTA) permits immediate replication of eQTLs using co-twins (93%–98%) and validation of the considerable gain in eQTL discovery after FA correction. We highlight the challenges of comparing eQTLs between tissues. After verifying previous significance threshold-based estimates of tissue-specificity, we show their limitations given their dependency on statistical power. We propose that continuous estimates of the proportion of tissue-shared signals and direct comparison of the magnitude of effect on the fold change in expression are essential properties that jointly provide a biologically realistic view of tissue-specificity. Under this framework we demonstrate that 30% of eQTLs are shared among the three tissues studied, while another 29% appear exclusively tissue-specific. However, even among the shared eQTLs, a substantial proportion (10%–20%) have significant differences in the magnitude of fold change between genotypic classes across tissues. Our results underline the need to account for the complexity of eQTL tissue-specificity in an effort to assess consequences of such variants for complex traits. INTRODUCTION Gene expression is an essential cellular function whose regulation determines a significant proportion of the phenotypic variance. Using microarrays and recently second generation sequencing (RNA-seq) [1], [2], major progress has been made in understanding the genetics of human gene expression and identifying loci that drive differential expression across individuals [3], [4], populations [5]–[7] and tissues [7]–[11]. This development is especially valuable for the biological analysis of genome-wide association (GWAS) signals [12], which often map to non-genic regions and are thus hard to interpret in the absence of additional information [13]. Transcript abundance is a very proximal endophenotype affected by genetic variation and has already facilitated the identification of candidate susceptibility genes for metabolic disease traits [14], asthma [15] or Crohn's disease [16]. This has been mostly possible when the tissue of expression was relevant to the interrogated complex trait, as disease phenotypes manifest themselves only in certain tissues. eQTLs discovered in LCLs have primarily helped explain GWAS associations with immunity-related disorders [17], [18] while associations with obesity-related traits were mostly observed when gene expression was quantified in adipose tissue [9]. Nevertheless, our guess of tissue relevance is yet far from satisfactory [19], reinforcing thus the incontestable value of measuring expression in multiple cell-types (including primary tissues reflecting in vivo patterns). Transcriptional regulatory networks are expected to dictate tissue-specificity of regulatory effects [20], but the extent of this is still under debate. Depending on the cell-types compared and the eQTL discovery methods used, current estimates for tissue-specificity of eQTLs range from ∼30% (liver, adipose tissue) [21] to 70–80% (LCL, fibroblasts, T cells) [7]. In this study we investigated various aspects of tissue-specificity and we emphasize the importance of accounting not only for statistical significance but also for continuous biological properties of regulatory variants, such as fold change in expression. We explored the complexity of the human cis-regulatory variation landscape in three tissues (LCL, skin and fat) derived from a subset of female Caucasian twins aged between 40 and 87 years old (mean 62 years) from the UK Adult Twin registry [22]. The present study represents the pilot phase of the MuTHER project (Multiple Tissue Human Expression Resource—http://www.muther.ac.uk/), a major resource initiated to enhance our knowledge about common trait susceptibility by providing genome-wide expression, methylation and eventually transcriptome sequencing information for 855 extensively phenotyped twins (clinical, anthropometric, life-style information as well as a wide range of biological measurements are available). RESULTS Gene expression was quantified in LCL, skin and fat using Illumina's whole genome expression array (HumanHT-12 version 3) containing 48,803 probes in three technical replicates [E-MTAB-522]. Log2 - transformed expression signals were normalized separately per tissue by quantile normalization across replicates followed by quantile normalization across individuals. 27,499 probes mapping uniquely to 18,170 Ensembl genes were retained for further analysis. The same individuals had also been genotyped with Illumina's 1M-Duo and 1.2M-Duo chips; 865,544 SNPs with MAF>1% passed quality check (QC). The overlapping set of successfully genotyped samples with available expression data amounted to 156 individuals for LCL (30 MZ pairs, 37 DZ pairs, 22 singletons), 160 for skin (31 MZ pairs, 37 DZ pairs, 24 singletons) and 166 for fat (31 MZ pairs, 40 DZ pairs, 24 singletons). This final dataset was used for eQTL analysis. We tested for SNP-gene expression associations (eQTLs) separately in each tissue. We considered only unrelated individuals at a time by separating twins from the same pair and thus performing two independent eQTL analyses per tissue. This study design, hereafter named Matched Co-Twin Analysis (MCTA), permits immediate replication and validation of eQTL discoveries. We used Spearman Rank Correlation (SRC) to detect associations and restricted our search to cis effects located within 1Mb on either side of a gene's transcription start site (TSS). Statistical significance was assessed at different thresholds using permutations (10,000 per gene) [5]. We detected an abundance of cis eQTLs (Table S1A) per tissue at a comparable rate to other studies of similar sample size [5], [7]. The reported eQTLs appear robust as they replicate well between individuals of the two co-twin groups per tissue. We measured the eQTL overlap in a continuous fashion by taking the significant SNP-gene associations from one co-twin set and estimating the proportion of true associations (π1 statistic [23], see Materials and Methods) on the distribution of corresponding p-values in the reciprocal co-twin validation set. High levels of eQTL replication were observed across co-twins, with a mean π1 of 0.93 in skin and 0.98 in LCL and fat (Table 1). We also measured the estimated proportion of true positives among the subset of genes that did not replicate in the co-twin at the same threshold. This too is high (π1 = 0.84 for skin and 0.94 for LCL and fat), suggesting that exact overlap of genes at a given permutation threshold (PT) is an underestimate of eQTL replication due to winner's curse. In other words, we detected eQTLs in the co-twin that clearly replicated the initial findings, but at p-values that marginally missed the initial discovery threshold. To further confirm the robustness of our discoveries, we overlapped the MuTHER LCL results with available eQTL data from two recent independent studies. 40% of the genes for which we detect LCL eQTLs overlap with eQTLs detected in HapMap 3 samples of European ancestry (CEU) (Stranger et al.