1
|
Hemani G, Knott S, Haley C. An evolutionary perspective on epistasis and the missing heritability. PLoS Genet 2013; 9:e1003295. [PMID: 23509438 PMCID: PMC3585114 DOI: 10.1371/journal.pgen.1003295] [Citation(s) in RCA: 95] [Impact Index Per Article: 8.6] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/21/2012] [Accepted: 12/17/2012] [Indexed: 01/04/2023] Open
Abstract
The relative importance between additive and non-additive genetic variance has been widely argued in quantitative genetics. By approaching this question from an evolutionary perspective we show that, while additive variance can be maintained under selection at a low level for some patterns of epistasis, the majority of the genetic variance that will persist is actually non-additive. We propose that one reason that the problem of the “missing heritability” arises is because the additive genetic variation that is estimated to be contributing to the variance of a trait will most likely be an artefact of the non-additive variance that can be maintained over evolutionary time. In addition, it can be shown that even a small reduction in linkage disequilibrium between causal variants and observed SNPs rapidly erodes estimates of epistatic variance, leading to an inflation in the perceived importance of additive effects. We demonstrate that the perception of independent additive effects comprising the majority of the genetic architecture of complex traits is biased upwards and that the search for causal variants in complex traits under selection is potentially underpowered by parameterising for additive effects alone. Given dense SNP panels the detection of causal variants through genome-wide association studies may be improved by searching for epistatic effects explicitly. In this study we have shown that two independent problems may have a common cause. Why do traits under selection exhibit additive genetic variance, and why is the proportion of the heritability explained by additive effects much smaller than the total heritability estimated to exist? Our results indicate that epistatic interactions can allow deleterious mutations to persist under selection and that these interactions can abate the depletion of additive genetic variation. Furthermore, a much larger element of non-additive genetic variance is maintained, which supports the notion that the heritability estimated from family studies could be a mixture of both additive and non-additive components. We show that searching directly for epistatic effects greatly improves the discovery of variants under selection, despite the multiple testing penalty being much larger. Finally, we demonstrate that common practices in genome-wide association studies could lead to both an ascertainment bias in detecting additive effects and a confirmation bias in perceiving that most of the genetic variance is additive.
Collapse
Affiliation(s)
- Gibran Hemani
- The Roslin Institute and Royal (Dick) School of Veterinary Science, University of Edinburgh, Edinburgh, United Kingdom
- * E-mail:
| | - Sara Knott
- MRC Human Genetics Unit, MRC Institute of Genetics and Molecular Medicine, University of Edinburgh, Western General Hospital, Edinburgh, United Kingdom
| | - Chris Haley
- The Roslin Institute and Royal (Dick) School of Veterinary Science, University of Edinburgh, Edinburgh, United Kingdom
- Institute for Evolutionary Biology, University of Edinburgh, Edinburgh, United Kingdom
| |
Collapse
|
2
|
Polanczyk M, Yellayi S, Zamora A, Subramanian S, Tovey M, Vandenbark AA, Offner H, Zachary JF, Fillmore PD, Blankenhorn EP, Gustafsson JA, Teuscher C. Estrogen receptor-1 (Esr1) and -2 (Esr2) regulate the severity of clinical experimental allergic encephalomyelitis in male mice. THE AMERICAN JOURNAL OF PATHOLOGY 2004; 164:1915-24. [PMID: 15161628 PMCID: PMC1615766 DOI: 10.1016/s0002-9440(10)63752-2] [Citation(s) in RCA: 32] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/28/2022]
Abstract
Estrogens and estrogen-receptor signaling function in establishing and regulating the female immune system and it is becoming increasingly evident that they may play a similar role in males. We report that B10.PL/SnJ male mice with a disrupted estrogen receptor-1 (alpha) gene (Esr1(-/-)) develop less severe clinical experimental allergic encephalomyelitis (EAE) compared to either Esr1(+/-) or wild-type (Esr1(+/+)) controls when immunized with myelin basic protein peptide Ac1-11 (MBP(Ac1-11)). In contrast, the disease course in B10.PL/SnJ male mice with a disrupted estrogen receptor-2 (beta) gene (Esr2(-/-)) does not differ from that of wild-type (Esr2(+/+)) mice. However, Esr2(+/-) mice do develop more severe clinical disease with an earlier onset indicating that heterosis at Esr2 plays a significant role in regulating EAE in males. No significant differences in central nervous system histopathology or MBP(Ac1-11)-specific T-cell responses as assessed by proliferation and interleukin-2 production were observed as a function of either Esr1 or Esr2 genotype. An analysis of cytokine/chemokine secretion by MBP(Ac1-11)-specific T cells revealed unique Esr1 and Esr2 genotype-dependent regulation. Interferon-gamma secretion was found to be negatively regulated by Esr1 whereas interleukin-6 and tumor necrosis factor-alpha secretion exhibited classical Esr2 gene dose responses. Interestingly, MCP-1 displayed distinctively unique patterns of genotype-dependent regulation by Esr1 and Esr2. The contribution of the hematopoietic and nonhematopoietic cellular compartments associated with the heterotic effect at Esr2 in regulating the severity of clinical EAE was identified using reciprocal hematopoietic radiation bone marrow chimeras generated between male wild-type and Esr2(+/-) mice. Wild-type --> Esr2(+/-) mice exhibited EAE equivalent in severity to that seen in Esr2(+/-) --> Esr2(+/-) control constructs; both of which were more severe than the clinical signs observed in Esr2(+/-) --> wild-type and wild-type --> wild-type mice. These results indicate that the heterotic effect at Esr2 is a function of the nonhematopoietic compartment.
Collapse
Affiliation(s)
- Magdalena Polanczyk
- Department of Neurology, Oregon Health and Science University, Portland, Oregon, USA
| | | | | | | | | | | | | | | | | | | | | | | |
Collapse
|
3
|
Abstract
Molecular heterosis occurs when subjects heterozygous for a specific genetic polymorphism show a significantly greater effect (positive heterosis) or lesser effect (negative heterosis) for a quantitative or dichotomous trait than subjects homozygous for either allele. At a molecular level heterosis appears counterintuitive to the expectation that if the 1 allele of a two-allele polymorphism is associated with a decrease in gene expression, those carrying the 11 genotype should show the greatest effect, 12 heterozygotes should be intermediate, and 22 homozygotes should show the least effect. We review the accumulating evidence that molecular heterosis is common in humans and may occur in up to 50% of all gene associations. A number of examples are reviewed, including those for the following genes: ADRA2C, C3 complement, DRD1, DRD2, DRD3, DRD4, ESR1, HP, HBB, HLA-DR DQ, HTR2A, properdin B, SLC6A4, PNMT, and secretor. Several examples are given in which the heterosis is gender-specific. Three explanations for molecular heterosis are proposed. The first is based on an inverted U-shaped response curve in which either to little or too much gene expression is deleterious, with optimal gene expression occurring in 12 heterozygotes. The second proposes an independent third factor causing a hidden stratification of the sample such that for in one set of subjects 11 homozygosity is associated with the highest phenotype score, while in the other set, 22 homozygosity is associated with the highest phenotype score. The third explanation suggests greater fitness in 12 heterozygotes because they show a broader range of gene expression than 11 or 22 homozygotes. Allele-based linkage techniques usually miss heterotic associations. Because up to 50% of association studies show a heterosis effect, this can significantly diminish the power of family-based linkage and association studies.
Collapse
Affiliation(s)
- D E Comings
- Department of Medical Genetics, City of Hope Medical Center, Duarte, California 91010, USA
| | | |
Collapse
|
4
|
Shine HD, Readhead C, Popko B, Hood L, Sidman RL. Morphometric Analysis of Normal, Mutant, and Transgenic CNS: Correlation of Myelin Basic Protein Expression to Myelinogenesis. J Neurochem 1992; 58:342-9. [PMID: 1370079 DOI: 10.1111/j.1471-4159.1992.tb09316.x] [Citation(s) in RCA: 86] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/28/2022]
Abstract
The neurological mutant mice shiverer (shi) and myelin deficient (shimld) lack a functional gene for the myelin basic proteins (MBP), have virtually no myelin in their CNS, shiver, seize, and die early. Mutant mice homozygous for an MBP transgene have MBP mRNA and MBP in net amounts approximately 25% of normal, have compact myelin, do not shiver or seize, and live normal life spans. We bred mice with various combinations of the normal, transgenic, shi, and shimld genes to produce mice that expressed MBP mRNA at levels of 0, 5, 12.5, 17.5, 50, 62.5, and 100% of normal. The CNS of these mice were analyzed for MBP content, tissue localization of MBP, degree of myelination, axon size, and myelin thickness. MBP protein content correlated with predicted MBP gene expression. Immunocytochemical staining localized MBP to white matter in normal and transgenic shi mice with an intensity of staining comparable to the degree of MBP gene expression. An increase in the percentage of myelinated axons and the thickness of myelin correlated with increased gene expression up to 50% of normal. The percentage of myelinated axons and myelin thickness remained constant at expression levels greater than 50%. The presence of axons loosely wrapped with oligodendrocytic membrane in mice expressing lower amounts of MBP mRNA and protein suggested that the oligodendroglia produced sufficient MBP to elicit axon wrapping but not enough to form compact myelin. Mean axon circumference of myelinated axons was greater than axon circumference of unmyelinated axons at each level of gene expression, further evidence that oligodendroglial cells preferentially myelinate axons of larger caliber.(ABSTRACT TRUNCATED AT 250 WORDS)
Collapse
Affiliation(s)
- H D Shine
- Center for Biotechnology, Baylor College of Medicine, Houston, Texas 77030
| | | | | | | | | |
Collapse
|
5
|
Abstract
Multiple sclerosis (MS) and a number of related distinctive diseases are characterized by the active degradation of central nervous system (CNS) myelin, an axonal sheath comprised essentially of proteins and lipids. These demyelinating diseases appear to arise from complex interactions of genetic, immunological, infective, and biochemical mechanisms. While circumstances of MS etiology remain hypothetical, one persistent theme involves recognition by the immune system of myelin-specific antigens derived from myelin basic protein (MBP), the most abundant extrinsic myelin membrane protein, and/or another equally susceptible myelin protein or lipid component. Knowledge of the biochemical and physical-chemical properties of myelin proteins and lipids, particularly their composition, organization, structure, and accessibility with respect to the compacted myelin multilayers, thus becomes central to the understanding of how and why these antigens become selected during the development of MS. This review focuses on current understanding of the molecular basis underlying demyelinating disease as it may relate to the impact of the various protein and lipid components on myelin morphology; the precise molecular architecture of this membrane as dictated by protein-lipid and lipid-lipid interactions; and the relationship, if any, between the protein/lipid components and the destruction of myelin in pathological situations.
Collapse
Affiliation(s)
- C M Deber
- Research Institute, Hospital for Sick Children, Toronto, Ontario, Canada
| | | |
Collapse
|
6
|
Kimura M, Sato M, Akatsuka A, Nozawa-Kimura S, Takahashi R, Yokoyama M, Nomura T, Katsuki M. Restoration of myelin formation by a single type of myelin basic protein in transgenic shiverer mice. Proc Natl Acad Sci U S A 1989; 86:5661-5. [PMID: 2473474 PMCID: PMC297684 DOI: 10.1073/pnas.86.14.5661] [Citation(s) in RCA: 79] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/01/2023] Open
Abstract
A minigene containing mouse cDNA coding for the smallest type of myelin basic protein and including the native promoter was constructed and used to produce transgenic shiverer mice. The hypomyelinating mouse, the shiverer, has a deletion in its myelin basic protein gene, lacks all four types of myelin basic protein in its myelin, and shows abnormal behavior such as violent tremors. Five of twenty-one transgenic shiverer mice showed recovered protein synthesis, compact myelin formation, and normal behavior. These results suggest that a single type of myelin basic protein restores myelin formation and returns the shivering phenotype to normal in the transgenic shiverer mouse.
Collapse
Affiliation(s)
- M Kimura
- Department of DNA Biology, School of Medicine, Tokai University, Kanagawa, Japan
| | | | | | | | | | | | | | | |
Collapse
|
7
|
Katsuki M, Sato M, Kimura M, Yokoyama M, Kobayashi K, Nomura T. Conversion of normal behavior to shiverer by myelin basic protein antisense cDNA in transgenic mice. Science 1988; 241:593-5. [PMID: 2456614 DOI: 10.1126/science.2456614] [Citation(s) in RCA: 158] [Impact Index Per Article: 4.4] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/01/2023]
Abstract
Myelin basic proteins (MBPs) are coded by the single gene necessary for myelin formation in the central nervous system of the mouse. An antisense MBP mini-gene was constructed and used to determine the function of antisense DNA in transgenic mice. Several transgenic offspring of a founder transgenic mouse, AS100, were converted from the normal to mutant shiverer phenotype. Antisense MBP messenger RNA was expressed in these mice, and the endogenous MBP messenger RNA, the MBP, and the myelination in the central nervous system were reduced.
Collapse
Affiliation(s)
- M Katsuki
- Department of DNA Biology, School of Medicine, Tokai University, Isehara, Japan
| | | | | | | | | | | |
Collapse
|
8
|
Abstract
We have recently described a murine model for studying aspects of myelination (Seyfried and Yu, 1980; Ebato et al., 1983; Miskimins et al., 1986). This mouse shows hypermyelination during the period of most active synthesis of myelin, 9 to 21 days post-natal. The myelin parameters showing an increase were all measured in the central nervous system. We investigated here whether this effect extends into the peripheral nervous system. Our results indicate that the hypermyelination is limited to the central nervous system.
Collapse
Affiliation(s)
- R Miskimins
- Department of Neurology, Yale University School of Medicine, New Haven, Connecticut 06510
| | | |
Collapse
|
9
|
Abstract
Previous studies have indicated that the brains of DBA/2J (D2) mice have a more heavily myelinated CNS than those of C57BL/6J (B6) at postnatal days 17-21. However, the amount of myelin in the brains of F1 (B6 X D2) hybrids is even higher than in their parental strains. To investigate further factors involved in regulating myelinogenesis in these mice, we have focused on the synthesis of cerebrosides and sulfatides, galactolipids enriched in myelin. Brain slices from 14-, 17-, and 21-day-old D2, B6, and F1 mice were incubated with [3H]galactose and [35S]sulfate. After incubation, microsomes, myelin, and oligodendroglial cells were isolated, and the galactolipids were analyzed. At 21 days of age, the labeling of cerebrosides in F1 mice was higher than in D2 and B6 mice when the results were expressed as microsomal or myelin radioactivity per gram wet weight. At 14 and 17 days of age, the labeling of cerebrosides in F1 animals was similar to that in D2 mice and was considerably higher than that in B6 mice. The labeling of sulfatides in F1 animals was significantly higher than in the B6 parent at all ages studied, whereas it remained higher than that in the D2 parent only at 17 days of age. A similar relationship among the strains was observed when the synthesis of myelin galactolipids was estimated by measuring the in vitro activity of UDP-galactose:ceramide galactosyltransferase and 3'-phosphoadenylyl sulfate:galactosylceramide 3'-sulfotransferase. The results indicate that the increased accumulation of myelin galactolipids previously reported in the F1 mice is partially due to enhanced synthetic activity.
Collapse
|
10
|
Miskimins R, Yu RK. Developmental expression of myelin basic protein mRNA in a hypermyelinating mouse. J Neurosci Res 1986; 16:219-25. [PMID: 2427737 DOI: 10.1002/jnr.490160119] [Citation(s) in RCA: 8] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/31/2022]
Abstract
Polysomal poly (A)+ RNA and total cellular RNA were isolated from the brains of C57BL/6J (B6), DBA/2J (D2), and F1 hybrid mice. Both types of RNA were isolated at 9, 12, 16, 21, and 30 days of age. By using a 32P-labelled cDNA probe for myelin basic protein (MBP; pMBP-1) the amounts of RNA coding for MBP present in the three strains at each age were determined. For both the polysomal poly(A)+ and total cellular RNA the peak of MBP RNA accumulation was at about 16-18 days for all three strains. These results imply that the developmental pattern of MBP RNA synthesis for F1 mice is not shifted with respect to the parental strains. For the polysomal poly(A)+ RNA, F1 had the greatest amount of MBP mRNA, B6 the least, and D2 an intermediate amount. F1 again had the greatest amount of MBP RNA from the total cellular pool at all ages. However, B6 had a greater amount of MBP RNA in the total cellular RNA pool than did D2. This demonstrates elevated transcription of the MBP gene in the F1 or an increase in message stability in this strain. It also suggests some form of translational control for MBP RNA in the B6 strain. In addition a second species of MBP RNA was found in the total cellular RNA pool. This species is larger than the major polysomal RNA band and may be a nuclear precursor.
Collapse
|