Applied statistical genetics

Aaron Day-Williams

adw@sanger.ac.ukPostdoctoral Fellow

Kalliope Panoutsopoulou

Staff Scientist

Lorraine Southam

Senior Staff Scientist

Ioanna Tachmazidou

it3@sanger.ac.ukStatistical Geneticist

Aaron Day-Williams

adw@sanger.ac.uk Postdoctoral Fellow

EDUCATION

Ph.D., Department of Human Genetics, University of California Los Angeles 2009

B.Sc., Microbiology and Cell Science, University of Florida 2000

RESEARCH EXPERIENCE

University of California Los Angeles, Department of Human Genetics, 2001-2009

Los Alamos National Laboratory, Bioinformatics/Database Software Engineer, Bioscience Division, 2000-2001

Los Alamos National Laboratory, Undergraduate Research Intern, Bioscience Division, 1999

The University of Florida, Undergraduate Research Assistant, Department of Microbiology, 1998-2000

Research

Investigating the role of rare variants in the biology of complex disease, and how to utilize them in genome-wide association analysis. Specifically addressing issues in designing and analyzing next-generation sequencing projects to discover rare variants relevant for disease.

References

• A variant in MCF2L is associated with osteoarthritis.

  Day-Williams AG, Southam L, Panoutsopoulou K, Rayner NW, Esko T, Estrada K, Helgadottir HT, Hofman A, Ingvarsson T, Jonsson H, Keis A, Kerkhof HJ, Thorleifsson G, Arden NK, Carr A, Chapman K, Deloukas P, Loughlin J, McCaskie A, Ollier WE, Ralston SH, Spector TD, Wallis GA, Wilkinson JM, Aslam N, Birell F, Carluke I, Joseph J, Rai A, Reed M, Walker K, arcOGEN Consortium, Doherty SA, Jonsdottir I, Maciewicz RA, Muir KR, Metspalu A, Rivadeneira F, Stefansson K, Styrkarsdottir U, Uitterlinden AG, van Meurs JB, Zhang W, Valdes AM, Doherty M and Zeggini E

  Wellcome Trust Sanger Institute, Hinxton, UK.

  Osteoarthritis (OA) is a prevalent, heritable degenerative joint disease with a substantial public health impact. We used a 1000-Genomes-Project-based imputation in a genome-wide association scan for osteoarthritis (3177 OA cases and 4894 controls) to detect a previously unidentified risk locus. We discovered a small disease-associated set of variants on chromosome 13. Through large-scale replication, we establish a robust association with SNPs in MCF2L (rs11842874, combined odds ratio [95% confidence interval] 1.17 [1.11-1.23], p = 2.1 × 10(-8)) across a total of 19,041 OA cases and 24,504 controls of European descent. This risk locus represents the third established signal for OA overall. MCF2L regulates a nerve growth factor (NGF), and treatment with a humanized monoclonal antibody against NGF is associated with reduction in pain and improvement in function for knee OA patients.

  American journal of human genetics 2011;89;3;446-50

  PUBMED: 21871595; PMC: 3169824; DOI: 10.1016/j.ajhg.2011.08.001

• Linkage analysis without defined pedigrees.

  Day-Williams AG, Blangero J, Dyer TD, Lange K and Sobel EM

  Department of Human Genetics, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA 90095-7088, USA.

  The need to collect accurate and complete pedigree information has been a drawback of family-based linkage and association studies. Even in case-control studies, investigators should be aware of, and condition on, familial relationships. In single nucleotide polymorphism (SNP) genome scans, relatedness can be directly inferred from the genetic data rather than determined through interviews. Various methods of estimating relatedness have previously been implemented, most notably in PLINK. We present new fast and accurate algorithms for estimating global and local kinship coefficients from dense SNP genotypes. These algorithms require only a single pass through the SNP genotype data. We also show that these estimates can be used to cluster individuals into pedigrees. With these estimates in hand, quantitative trait locus linkage analysis proceeds via traditional variance components methods without any prior relationship information. We demonstrate the success of our algorithms on simulated and real data sets. Our procedures make linkage analysis as easy as a typical genomewide association study.

  Funded by: NIGMS NIH HHS: GM053275; NIMH NIH HHS: MH059490

  Genetic epidemiology 2011;35;5;360-70

  PUBMED: 21465549; DOI: 10.1002/gepi.20584

• Insights into the genetic architecture of osteoarthritis from stage 1 of the arcOGEN study.

  Panoutsopoulou K, Southam L, Elliott KS, Wrayner N, Zhai G, Beazley C, Thorleifsson G, Arden NK, Carr A, Chapman K, Deloukas P, Doherty M, McCaskie A, Ollier WE, Ralston SH, Spector TD, Valdes AM, Wallis GA, Wilkinson JM, Arden E, Battley K, Blackburn H, Blanco FJ, Bumpstead S, Cupples LA, Day-Williams AG, Dixon K, Doherty SA, Esko T, Evangelou E, Felson D, Gomez-Reino JJ, Gonzalez A, Gordon A, Gwilliam R, Halldorsson BV, Hauksson VB, Hofman A, Hunt SE, Ioannidis JP, Ingvarsson T, Jonsdottir I, Jonsson H, Keen R, Kerkhof HJ, Kloppenburg MG, Koller N, Lakenberg N, Lane NE, Lee AT, Metspalu A, Meulenbelt I, Nevitt MC, O'Neill F, Parimi N, Potter SC, Rego-Perez I, Riancho JA, Sherburn K, Slagboom PE, Stefansson K, Styrkarsdottir U, Sumillera M, Swift D, Thorsteinsdottir U, Tsezou A, Uitterlinden AG, van Meurs JB, Watkins B, Wheeler M, Mitchell S, Zhu Y, Zmuda JM, arcOGEN Consortium, Zeggini E and Loughlin J

  Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK.

  Objectives: The genetic aetiology of osteoarthritis has not yet been elucidated. To enable a well-powered genome-wide association study (GWAS) for osteoarthritis, the authors have formed the arcOGEN Consortium, a UK-wide collaborative effort aiming to scan genome-wide over 7500 osteoarthritis cases in a two-stage genome-wide association scan. Here the authors report the findings of the stage 1 interim analysis.

  Methods: The authors have performed a genome-wide association scan for knee and hip osteoarthritis in 3177 cases and 4894 population-based controls from the UK. Replication of promising signals was carried out in silico in five further scans (44,449 individuals), and de novo in 14 534 independent samples, all of European descent.

  Results: None of the association signals the authors identified reach genome-wide levels of statistical significance, therefore stressing the need for corroboration in sample sets of a larger size. Application of analytical approaches to examine the allelic architecture of disease to the stage 1 genome-wide association scan data suggests that osteoarthritis is a highly polygenic disease with multiple risk variants conferring small effects.

  Conclusions: Identifying loci conferring susceptibility to osteoarthritis will require large-scale sample sizes and well-defined phenotypes to minimise heterogeneity.

  Funded by: Arthritis Research UK

  Annals of the rheumatic diseases 2011;70;5;864-7

  PUBMED: 21177295; PMC: 3070286; DOI: 10.1136/ard.2010.141473

• The effect of next-generation sequencing technology on complex trait research.

  Day-Williams AG and Zeggini E

  Wellcome Trust Sanger Institute, Hinxton, Cambridge, UK.

  Background: Advances in the understanding of complex trait genetics have always been enabled by advances in genomic technology. Next-generation sequencing (NGS) is set to revolutionize the way complex trait genetics research is carried out.

  Results: NGS has multiple applications in the field of human genetics, but is accompanied by substantial study design, analysis and interpretation challenges. This review discusses key aspects of study design considerations, data handling issues and required analytical developments. We also highlight early successes in mapping genetic traits using NGS.

  Conclusion: NGS opens the entire spectrum of genomic alterations for the genetic analysis of complex traits and there are early publications illustrating its power. Continuing development in analytical tools will allow the promise of NGS to be realized.

  European journal of clinical investigation 2011;41;5;561-7

  PUBMED: 21155765; PMC: 3107422; DOI: 10.1111/j.1365-2362.2010.02437.x

• An evaluation of different target enrichment methods in pooled sequencing designs for complex disease association studies.

  Day-Williams AG, McLay K, Drury E, Edkins S, Coffey AJ, Palotie A and Zeggini E

  Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire, United Kingdom.

  Pooled sequencing can be a cost-effective approach to disease variant discovery, but its applicability in association studies remains unclear. We compare sequence enrichment methods coupled to next-generation sequencing in non-indexed pools of 1, 2, 10, 20 and 50 individuals and assess their ability to discover variants and to estimate their allele frequencies. We find that pooled resequencing is most usefully applied as a variant discovery tool due to limitations in estimating allele frequency with high enough accuracy for association studies, and that in-solution hybrid-capture performs best among the enrichment methods examined regardless of pool size.

  Funded by: Wellcome Trust: WT088885/Z/09/Z

  PloS one 2011;6;11;e26279

  PUBMED: 22069447; PMC: 3206031; DOI: 10.1371/journal.pone.0026279

• Unifying ideas for non-parametric linkage analysis.

  Day-Williams AG, Blangero J, Dyer TD, Lange K and Sobel EM

  Department of Human Genetics, David Geffen School of Medicine, University of California at Los Angeles, USA. adw @ sanger.ac.uk

  Objectives: Non-parametric linkage analysis (NPL) exploits marker allele sharing among affected relatives to map genes influencing complex traits. Computational barriers force approximate analysis on large pedigrees and the adoption of a questionable perfect data assumption (PDA) in assigning p values. To improve NPL significance testing on large pedigrees, we examine the adverse consequences of missing data and PDA. We also introduce a novel statistic, Q-NPL, appropriate for NPL analysis of quantitative traits.

  Methods: Using simulated and real data sets with qualitative traits, we compare NPL analysis results for four testing procedures and various degrees of missing data. The simulated data sets vary from all nuclear families, to all large pedigrees, to a mix of pedigrees of different sizes. We implemented the Kong and Cox linear adjustment of p values in the software packages Mendel and SimWalk. We perform similar analysis with Q-NPL on quantitative traits of various heritabilities.

  Results: The Kong and Cox extension for significance testing is robust to realistic missing data patterns, greatly improves p values in approximate analyses, and works equally well for qualitative and quantitative traits and small and large pedigrees. The Q-NPL statistic is robust to missing data and shows good power to detect linkage for quantitative traits with a wide spectrum of heritabilities.

  Conclusions: The Kong and Cox extension should be a standard tool for calculating NPL p values. It allows the combination of exact and estimated analyses into a single significance score. Q-NPL should be a standard statistic for NPL analysis of quantitative traits. The new statistics are implemented in Mendel and SimWalk.

  Human heredity 2011;71;4;267-80

  PUBMED: 21822022; DOI: 10.1159/000323752

• CCRaVAT and QuTie-enabling analysis of rare variants in large-scale case control and quantitative trait association studies.

  Lawrence R, Day-Williams AG, Elliott KS, Morris AP and Zeggini E

  Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK.

  Background: Genome-wide association studies have been successful in finding common variants influencing common traits. However, these associations only account for a fraction of trait heritability. There has been a shift in the field towards studying low frequency and rare variants, which are now widely recognised as putative complex trait determinants. Despite this increasing focus on examining the role of low frequency and rare variants in complex disease susceptibility, there is a lack of user-friendly analytical packages implementing powerful association tests for the analysis of rare variants.

  Results: We have developed two software tools, CCRaVAT (Case-Control Rare Variant Analysis Tool) and QuTie (Quantitative Trait), which enable efficient large-scale analysis of low frequency and rare variants. Both programs implement a collapsing method examining the accumulation of low frequency and rare variants across a locus of interest that has more power than single variant analysis. CCRaVAT carries out case-control analyses whereas QuTie has been developed for continuous trait analysis.

  Conclusions: CCRaVAT and QuTie are easy to use software tools that allow users to perform genome-wide association analysis on low frequency and rare variants for both binary and quantitative traits. The software is freely available and provides the genetics community with a resource to perform association analysis on rarer genetic variants.

  Funded by: Wellcome Trust: 079557, 079557MA, 081682, WT088885/Z/09/Z

  BMC bioinformatics 2010;11;527

  PUBMED: 20964851; PMC: 2973964; DOI: 10.1186/1471-2105-11-527

• Large-scale association analysis of TNF/LTA gene region polymorphisms in type 2 diabetes.

  Boraska V, Rayner NW, Groves CJ, Frayling TM, Diakite M, Rockett KA, Kwiatkowski DP, Day-Williams AG, McCarthy MI and Zeggini E

  Department of Medical Biology, University of Split School of Medicine, Split, Croatia. vboraska@mefst.hr

  Background: The TNF/LTA locus has been a long-standing T2D candidate gene. Several studies have examined association of TNF/LTA SNPs with T2D but the majority have been small-scale and produced no convincing evidence of association. The purpose of this study is to examine T2D association of tag SNPs in the TNF/LTA region capturing the majority of common variation in a large-scale sample set of UK/Irish origin.

  Methods: This study comprised a case-control (1520 cases and 2570 control samples) and a family-based component (423 parent-offspring trios). Eleven tag SNPs (rs928815, rs909253, rs746868, rs1041981 (T60N), rs1800750, rs1800629 (G-308A), rs361525 (G-238A), rs3093662, rs3093664, rs3093665, and rs3093668) were selected across the TNF/LTA locus and genotyped using a fluorescence-based competitive allele specific assay. Quality control of the obtained genotypes was performed prior to single- and multi-point association analyses under the additive model.

  Results: We did not find any consistent SNP associations with T2D in the case-control or family-based datasets.

  Conclusions: The present study, designed to analyse a set of tag SNPs specifically selected to capture the majority of common variation in the TNF/LTA gene region, found no robust evidence for association with T2D. To investigate the presence of smaller effects of TNF/LTA gene variation with T2D, a large-scale meta-analysis will be required.

  Funded by: Medical Research Council: G0000934; Wellcome Trust: 068545/Z/02, 076113, WT088885/Z/09/Z

  BMC medical genetics 2010;11;69

  PUBMED: 20459604; PMC: 2873325; DOI: 10.1186/1471-2350-11-69

• GLIDERS--a web-based search engine for genome-wide linkage disequilibrium between HapMap SNPs.

  Lawrence R, Day-Williams AG, Mott R, Broxholme J, Cardon LR and Zeggini E

  Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK. z.r.yang@ex.ac.uk

  Background: A number of tools for the examination of linkage disequilibrium (LD) patterns between nearby alleles exist, but none are available for quickly and easily investigating LD at longer ranges (>500 kb). We have developed a web-based query tool (GLIDERS: Genome-wide LInkage DisEquilibrium Repository and Search engine) that enables the retrieval of pairwise associations with r2 >or= 0.3 across the human genome for any SNP genotyped within HapMap phase 2 and 3, regardless of distance between the markers.

  Description: GLIDERS is an easy to use web tool that only requires the user to enter rs numbers of SNPs they want to retrieve genome-wide LD for (both nearby and long-range). The intuitive web interface handles both manual entry of SNP IDs as well as allowing users to upload files of SNP IDs. The user can limit the resulting inter SNP associations with easy to use menu options. These include MAF limit (5-45%), distance limits between SNPs (minimum and maximum), r2 (0.3 to 1), HapMap population sample (CEU, YRI and JPT+CHB combined) and HapMap build/release. All resulting genome-wide inter-SNP associations are displayed on a single output page, which has a link to a downloadable tab delimited text file.

  Conclusion: GLIDERS is a quick and easy way to retrieve genome-wide inter-SNP associations and to explore LD patterns for any number of SNPs of interest. GLIDERS can be useful in identifying SNPs with long-range LD. This can highlight mis-mapping or other potential association signal localisation problems.

  Funded by: Wellcome Trust: 079557, 079557MA, 088885/Z/09/Z

  BMC bioinformatics 2009;10;367

  PUBMED: 19878600; PMC: 2777181; DOI: 10.1186/1471-2105-10-367

• Locus for quantitative HDL-cholesterol on chromosome 10q in Finnish families with dyslipidemia.

  Lilja HE, Suviolahti E, Soro-Paavonen A, Hiekkalinna T, Day A, Lange K, Sobel E, Taskinen MR, Peltonen L, Perola M and Pajukanta P

  Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland.

  Decreased HDL-cholesterol (HDL-C) and familial combined hyperlipidemia (FCHL) are the two most common familial dyslipidemias predisposing to premature coronary heart disease (CHD). These dyslipidemias share many phenotypic features, suggesting a partially overlapping molecular pathogenesis. This was supported by our previous pooled data analysis of the genome scans for low HDL-C and FCHL, which identified three shared chromosomal regions for a qualitative HDL-C trait on 8q23.1, 16q23.3, and 20q13.32. This study further investigates these regions as well as two other loci we identified earlier for premature CHD on 2q31 and Xq24 and a locus for high serum triglycerides (TGs) on 10q11. We analyzed 67 microsatellite markers in an extended study sample of 1,109 individuals from 92 low HDL-C or FCHL families using both qualitative and quantitative lipid phenotypes. These analyses provided evidence for linkage (a logarithm of odds score of 3.2) on 10q11 using a quantitative HDL-C trait. Importantly, this region, previously linked to TGs, body mass index, and obesity, provided evidence for association for quantitative TGs (P = 0.0006) and for a combined trait of HDL-C and TGs (P = 0.008) with marker D10S546. Suggestive evidence for linkage also emerged for HDL-C on 2q31 and for TGs on 20q13.32. Finnish families ascertained for dyslipidemias thus suggest that 10q11, 2q31, and 20q13.32 harbor loci for HDL-C and TGs.

  Funded by: NHLBI NIH HHS: HL 28481, HL 70150

  Journal of lipid research 2004;45;10;1876-84

  PUBMED: 15258200; DOI: 10.1194/jlr.M400141-JLR200

Kalliope Panoutsopoulou

- Staff Scientist

2008 - 2010 Postdoctoral Research Fellow, Wellcome Trust Centre for Human Genetics, Oxford, UK & Wellcome Trust Sanger Institute, Hinxton, UK

2002 - 2005 Research Associate, UMIST, Manchester, UK

2002 PhD in Genetics, University of Manchester, UK

1998 BSc (Hons) in Biochemistry and Applied Molecular Biology, UMIST, Manchester, UK

Research

We are part of the arcOGEN Consortium, a UK-wide collaboration and have carried out the largest genome-wide association scan of osteoarthritis. Our research has led to the identification of several novel loci implicated in the disease.

References

• Examining the overlap between genome-wide rare variant association signals and linkage peaks in rheumatoid arthritis.

  Eyre S, Ke X, Lawrence R, Bowes J, Panoutsopoulou K, Barton A, Thomson W, Worthington J and Zeggini E

  University of Manchester, Manchester, UK.

  Objective: With the exception of the major histocompatibility complex (MHC) and STAT4, no other rheumatoid arthritis (RA) linkage peak has been successfully fine-mapped to date. This apparent failure to identify association under peaks of linkage could be ascribed to the examination of common variation, when linkage is likely to be driven by rare variants. The purpose of this study was to investigate the overlap between genome-wide rare variant RA association signals observed in the Wellcome Trust Case Control Consortium (WTCCC) study and 11 replicating RA linkage peaks, defined as regions with evidence for linkage in >1 study.

  Methods: The WTCCC data set contained 40,482 variants with minor allele frequency of ≤0.05 in 1,860 RA patients and 2,938 controls. Genotypes of all rare variants within a given gene region were collapsed into a single locus and a global P value was calculated per gene.

  Results: The distribution of rare variant signals (association P≤10(-5)) was found to differ significantly between regions with and without linkage evidence (P=2×10(-17) by Fisher's exact test). No significant difference was observed after data from the MHC region were removed or when the effect of the HLA-DRB1 locus was accounted for.

  Conclusion: The results suggest that rare variant association signals are significantly overrepresented under linkage peaks in RA, but the effect is driven by the MHC. This is the first study to examine the overlap between linkage peaks and rare variant association signals genome-wide in a complex disease.

  Funded by: Arthritis Research UK: 18030; Wellcome Trust: 076113, 079557MA, WT088885/Z/09/Z

  Arthritis and rheumatism 2011;63;6;1522-6

  PUBMED: 21370227; DOI: 10.1002/art.30315

• Insights into the genetic architecture of osteoarthritis from stage 1 of the arcOGEN study.

  Panoutsopoulou K, Southam L, Elliott KS, Wrayner N, Zhai G, Beazley C, Thorleifsson G, Arden NK, Carr A, Chapman K, Deloukas P, Doherty M, McCaskie A, Ollier WE, Ralston SH, Spector TD, Valdes AM, Wallis GA, Wilkinson JM, Arden E, Battley K, Blackburn H, Blanco FJ, Bumpstead S, Cupples LA, Day-Williams AG, Dixon K, Doherty SA, Esko T, Evangelou E, Felson D, Gomez-Reino JJ, Gonzalez A, Gordon A, Gwilliam R, Halldorsson BV, Hauksson VB, Hofman A, Hunt SE, Ioannidis JP, Ingvarsson T, Jonsdottir I, Jonsson H, Keen R, Kerkhof HJ, Kloppenburg MG, Koller N, Lakenberg N, Lane NE, Lee AT, Metspalu A, Meulenbelt I, Nevitt MC, O'Neill F, Parimi N, Potter SC, Rego-Perez I, Riancho JA, Sherburn K, Slagboom PE, Stefansson K, Styrkarsdottir U, Sumillera M, Swift D, Thorsteinsdottir U, Tsezou A, Uitterlinden AG, van Meurs JB, Watkins B, Wheeler M, Mitchell S, Zhu Y, Zmuda JM, arcOGEN Consortium, Zeggini E and Loughlin J

  Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK.

  Objectives: The genetic aetiology of osteoarthritis has not yet been elucidated. To enable a well-powered genome-wide association study (GWAS) for osteoarthritis, the authors have formed the arcOGEN Consortium, a UK-wide collaborative effort aiming to scan genome-wide over 7500 osteoarthritis cases in a two-stage genome-wide association scan. Here the authors report the findings of the stage 1 interim analysis.

  Methods: The authors have performed a genome-wide association scan for knee and hip osteoarthritis in 3177 cases and 4894 population-based controls from the UK. Replication of promising signals was carried out in silico in five further scans (44,449 individuals), and de novo in 14 534 independent samples, all of European descent.

  Results: None of the association signals the authors identified reach genome-wide levels of statistical significance, therefore stressing the need for corroboration in sample sets of a larger size. Application of analytical approaches to examine the allelic architecture of disease to the stage 1 genome-wide association scan data suggests that osteoarthritis is a highly polygenic disease with multiple risk variants conferring small effects.

  Conclusions: Identifying loci conferring susceptibility to osteoarthritis will require large-scale sample sizes and well-defined phenotypes to minimise heterogeneity.

  Funded by: Arthritis Research UK

  Annals of the rheumatic diseases 2011;70;5;864-7

  PUBMED: 21177295; PMC: 3070286; DOI: 10.1136/ard.2010.141473

• The effect of genome-wide association scan quality control on imputation outcome for common variants.

  Southam L, Panoutsopoulou K, Rayner NW, Chapman K, Durrant C, Ferreira T, Arden N, Carr A, Deloukas P, Doherty M, Loughlin J, McCaskie A, Ollier WE, Ralston S, Spector TD, Valdes AM, Wallis GA, Wilkinson JM, arcOGEN consortium, Marchini J and Zeggini E

  Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK.

  Imputation is an extremely valuable tool in conducting and synthesising genome-wide association studies (GWASs). Directly typed SNP quality control (QC) is thought to affect imputation quality. It is, therefore, common practise to use quality-controlled (QCed) data as an input for imputing genotypes. This study aims to determine the effect of commonly applied QC steps on imputation outcomes. We performed several iterations of imputing SNPs across chromosome 22 in a dataset consisting of 3177 samples with Illumina 610 k (Illumina, San Diego, CA, USA) GWAS data, applying different QC steps each time. The imputed genotypes were compared with the directly typed genotypes. In addition, we investigated the correlation between alternatively QCed data. We also applied a series of post-imputation QC steps balancing elimination of poorly imputed SNPs and information loss. We found that the difference between the unQCed data and the fully QCed data on imputation outcome was minimal. Our study shows that imputation of common variants is generally very accurate and robust to GWAS QC, which is not a major factor affecting imputation outcome. A minority of common-frequency SNPs with particular properties cannot be accurately imputed regardless of QC stringency. These findings may not generalise to the imputation of low frequency and rare variants.

  Funded by: Arthritis Research UK: 18030; Wellcome Trust: 079557, 088885, WT079557MA, WT088885/Z/09/Z

  European journal of human genetics : EJHG 2011;19;5;610-4

  PUBMED: 21267008; PMC: 3083623; DOI: 10.1038/ejhg.2010.242

• Meta-analysis of genome-wide association studies confirms a susceptibility locus for knee osteoarthritis on chromosome 7q22.

  Evangelou E, Valdes AM, Kerkhof HJ, Styrkarsdottir U, Zhu Y, Meulenbelt I, Lories RJ, Karassa FB, Tylzanowski P, Bos SD, arcOGEN Consortium, Akune T, Arden NK, Carr A, Chapman K, Cupples LA, Dai J, Deloukas P, Doherty M, Doherty S, Engstrom G, Gonzalez A, Halldorsson BV, Hammond CL, Hart DJ, Helgadottir H, Hofman A, Ikegawa S, Ingvarsson T, Jiang Q, Jonsson H, Kaprio J, Kawaguchi H, Kisand K, Kloppenburg M, Kujala UM, Lohmander LS, Loughlin J, Luyten FP, Mabuchi A, McCaskie A, Nakajima M, Nilsson PM, Nishida N, Ollier WE, Panoutsopoulou K, van de Putte T, Ralston SH, Rivadeneira F, Saarela J, Schulte-Merker S, Shi D, Slagboom PE, Sudo A, Tamm A, Tamm A, Thorleifsson G, Thorsteinsdottir U, Tsezou A, Wallis GA, Wilkinson JM, Yoshimura N, Zeggini E, Zhai G, Zhang F, Jonsdottir I, Uitterlinden AG, Felson DT, van Meurs JB, Stefansson K, Ioannidis JP, Spector TD and Translation Research in Europe Applied Technologies for Osteoarthritis (TreatOA)

  Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Ioannina, Greece.

  Objectives: Osteoarthritis (OA) is the most prevalent form of arthritis and accounts for substantial morbidity and disability, particularly in older people. It is characterised by changes in joint structure, including degeneration of the articular cartilage, and its aetiology is multifactorial with a strong postulated genetic component.

  Methods: A meta-analysis was performed of four genome-wide association (GWA) studies of 2371 cases of knee OA and 35 909 controls in Caucasian populations. Replication of the top hits was attempted with data from 10 additional replication datasets.

  Results: With a cumulative sample size of 6709 cases and 44 439 controls, one genome-wide significant locus was identified on chromosome 7q22 for knee OA (rs4730250, p=9.2 × 10⁻⁹), thereby confirming its role as a susceptibility locus for OA.

  Conclusion: The associated signal is located within a large (500 kb) linkage disequilibrium block that contains six genes: PRKAR2B (protein kinase, cAMP-dependent, regulatory, type II, β), HPB1 (HMG-box transcription factor 1), COG5 (component of oligomeric golgi complex 5), GPR22 (G protein-coupled receptor 22), DUS4L (dihydrouridine synthase 4-like) and BCAP29 (B cell receptor-associated protein 29). Gene expression analyses of the (six) genes in primary cells derived from different joint tissues confirmed expression of all the genes in the joint environment.

  Funded by: Medical Research Council: G0000934; Wellcome Trust: 068545, WT079557MA, WT088885/Z/09/Z

  Annals of the rheumatic diseases 2011;70;2;349-55

  PUBMED: 21068099; DOI: 10.1136/ard.2010.132787

• Rare variation at the TNFAIP3 locus and susceptibility to rheumatoid arthritis.

  Bowes J, Lawrence R, Eyre S, Panoutsopoulou K, Orozco G, Elliott KS, Ke X, Morris AP, UKRAG, Thomson W, Worthington J, Barton A and Zeggini E

  Arthritis Research UK, Epidemiology Unit, University of Manchester, Manchester, UK.

  Genome-wide association studies (GWAS) conducted using commercial single nucleotide polymorphisms (SNP) arrays have proven to be a powerful tool for the detection of common disease susceptibility variants. However, their utility for the detection of lower frequency variants is yet to be practically investigated. Here we describe the application of a rare variant collapsing method to a large genome-wide SNP dataset, the Wellcome Trust Case Control Consortium rheumatoid arthritis (RA) GWAS. We partitioned the data into gene-centric bins and collapsed genotypes of low frequency variants (defined here as MAF ≤ 0.05) into a single count coupled with univariate analysis. We then prioritized gene regions for further investigation in an independent cohort of 3,355 cases and 2,427 controls based on rare variant signal p value and prior evidence to support involvement in RA. A total of 14,536 gene bins were investigated in the primary analysis and signals mapping to the TNFAIP3 and chr17q24 loci were selected for further investigation. We detected replicating association to low frequency variants in the TNFAIP3 gene (combined p = 6.6 × 10(-6)). Even though rare variants are not well-represented and can be difficult to genotype in GWAS, our study supports the application of low frequency variant collapsing methods to genome-wide SNP datasets as a means of exploiting data that are routinely ignored.

  Funded by: Arthritis Research UK: 17552; Wellcome Trust: 081682

  Human genetics 2010;128;6;627-33

  PUBMED: 20852893; PMC: 2978888; DOI: 10.1007/s00439-010-0889-1

• Finding common susceptibility variants for complex disease: past, present and future.

  Panoutsopoulou K and Zeggini E

  Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge, UK.

  The identification of complex disease susceptibility loci has been accelerated considerably by advances in high-throughput genotyping technologies, improved insight into correlation patterns of common variants and the availability of large-scale sample sets. Linkage scans and small-scale candidate gene studies have now given way to genome-wide association scans. In this review, we summarize insights gained from the past, highlight practical issues relating to the design and analysis of current state-of-the-art GWA studies and look into future trends in the field of human complex trait genetics.

  Funded by: Wellcome Trust: WT088885/Z/09/Z

  Briefings in functional genomics & proteomics 2009;8;5;345-52

  PUBMED: 19571035; PMC: 2758134; DOI: 10.1093/bfgp/elp020

Lorraine Southam

- Senior Staff Scientist

2001 - 2010 Research Assistant, University of Oxford, Botnar Research Centre, Oxford, UK

2000 - 2001 Research Assistant, University of Oxford, Institute of Molecular Medicine, Oxford, UK

1999 - 2000 Scientific Officer, Medical Research Council, Harwell, Didcot, Oxfordshire, UK

1996 - 1999 Research Assistant, University of Oxford, Wellcome Trust Centre for Human Genetics, Oxford, UK

1995 - 1996 Medical Laboratory Scientific Officer 1, University of Oxford, Wellcome Trust Centre for Human Genetics, Oxford, UK

1992 - 1995 Bsc (Hons) First class in Biomedical Science, Bradford University, Bradford, UK

Research

I form part of the arcOGEN Consortium analysis team aiming to identifying susceptibility loci that give rise to osteoarthritis. I am also involved in the UK10K and 500 exomes studies.

References

• Insights into the genetic architecture of osteoarthritis from stage 1 of the arcOGEN study.

  Panoutsopoulou K, Southam L, Elliott KS, Wrayner N, Zhai G, Beazley C, Thorleifsson G, Arden NK, Carr A, Chapman K, Deloukas P, Doherty M, McCaskie A, Ollier WE, Ralston SH, Spector TD, Valdes AM, Wallis GA, Wilkinson JM, Arden E, Battley K, Blackburn H, Blanco FJ, Bumpstead S, Cupples LA, Day-Williams AG, Dixon K, Doherty SA, Esko T, Evangelou E, Felson D, Gomez-Reino JJ, Gonzalez A, Gordon A, Gwilliam R, Halldorsson BV, Hauksson VB, Hofman A, Hunt SE, Ioannidis JP, Ingvarsson T, Jonsdottir I, Jonsson H, Keen R, Kerkhof HJ, Kloppenburg MG, Koller N, Lakenberg N, Lane NE, Lee AT, Metspalu A, Meulenbelt I, Nevitt MC, O'Neill F, Parimi N, Potter SC, Rego-Perez I, Riancho JA, Sherburn K, Slagboom PE, Stefansson K, Styrkarsdottir U, Sumillera M, Swift D, Thorsteinsdottir U, Tsezou A, Uitterlinden AG, van Meurs JB, Watkins B, Wheeler M, Mitchell S, Zhu Y, Zmuda JM, arcOGEN Consortium, Zeggini E and Loughlin J

  Wellcome Trust Sanger Institute, Hinxton, Cambridgeshire, UK.

  Objectives: The genetic aetiology of osteoarthritis has not yet been elucidated. To enable a well-powered genome-wide association study (GWAS) for osteoarthritis, the authors have formed the arcOGEN Consortium, a UK-wide collaborative effort aiming to scan genome-wide over 7500 osteoarthritis cases in a two-stage genome-wide association scan. Here the authors report the findings of the stage 1 interim analysis.

  Methods: The authors have performed a genome-wide association scan for knee and hip osteoarthritis in 3177 cases and 4894 population-based controls from the UK. Replication of promising signals was carried out in silico in five further scans (44,449 individuals), and de novo in 14 534 independent samples, all of European descent.

  Results: None of the association signals the authors identified reach genome-wide levels of statistical significance, therefore stressing the need for corroboration in sample sets of a larger size. Application of analytical approaches to examine the allelic architecture of disease to the stage 1 genome-wide association scan data suggests that osteoarthritis is a highly polygenic disease with multiple risk variants conferring small effects.

  Conclusions: Identifying loci conferring susceptibility to osteoarthritis will require large-scale sample sizes and well-defined phenotypes to minimise heterogeneity.

  Funded by: Arthritis Research UK

  Annals of the rheumatic diseases 2011;70;5;864-7

  PUBMED: 21177295; PMC: 3070286; DOI: 10.1136/ard.2010.141473

• The effect of genome-wide association scan quality control on imputation outcome for common variants.

  Southam L, Panoutsopoulou K, Rayner NW, Chapman K, Durrant C, Ferreira T, Arden N, Carr A, Deloukas P, Doherty M, Loughlin J, McCaskie A, Ollier WE, Ralston S, Spector TD, Valdes AM, Wallis GA, Wilkinson JM, arcOGEN consortium, Marchini J and Zeggini E

  Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK.

  Imputation is an extremely valuable tool in conducting and synthesising genome-wide association studies (GWASs). Directly typed SNP quality control (QC) is thought to affect imputation quality. It is, therefore, common practise to use quality-controlled (QCed) data as an input for imputing genotypes. This study aims to determine the effect of commonly applied QC steps on imputation outcomes. We performed several iterations of imputing SNPs across chromosome 22 in a dataset consisting of 3177 samples with Illumina 610 k (Illumina, San Diego, CA, USA) GWAS data, applying different QC steps each time. The imputed genotypes were compared with the directly typed genotypes. In addition, we investigated the correlation between alternatively QCed data. We also applied a series of post-imputation QC steps balancing elimination of poorly imputed SNPs and information loss. We found that the difference between the unQCed data and the fully QCed data on imputation outcome was minimal. Our study shows that imputation of common variants is generally very accurate and robust to GWAS QC, which is not a major factor affecting imputation outcome. A minority of common-frequency SNPs with particular properties cannot be accurately imputed regardless of QC stringency. These findings may not generalise to the imputation of low frequency and rare variants.

  Funded by: Arthritis Research UK: 18030; Wellcome Trust: 079557, 088885, WT079557MA, WT088885/Z/09/Z

  European journal of human genetics : EJHG 2011;19;5;610-4

  PUBMED: 21267008; PMC: 3083623; DOI: 10.1038/ejhg.2010.242

• Identification and characterization of novel parathyroid-specific transcription factor Glial Cells Missing Homolog B (GCMB) mutations in eight families with autosomal recessive hypoparathyroidism.

  Bowl MR, Mirczuk SM, Grigorieva IV, Piret SE, Cranston T, Southam L, Allgrove J, Bahl S, Brain C, Loughlin J, Mughal Z, Ryan F, Shaw N, Thakker YV, Tiosano D, Nesbit MA and Thakker RV

  Academic Endocrine Unit, Nuffield Department of Clinical Medicine, Oxford Centre for Diabetes, Endocrinology and Metabolism, University of Oxford, Churchill Hospital, Headington, Oxford OX3 7LJ, UK.

  GCMB is a member of the small transcription factor family GCM (glial cells missing), which are important regulators of development, present in vertebrates and some invertebrates. In man, GCMB encodes a 506 amino acid parathyroid gland-specific protein, mutations of which have been reported to cause both autosomal dominant and autosomal recessive hypoparathyroidism. We ascertained 18 affected individuals from 12 families with autosomal recessive hypoparathyroidism and have investigated them for GCMB abnormalities. Four different homozygous germline mutations were identified in eight families that originate from the Indian Subcontinent. These consisted of a novel nonsense mutation R39X; a missense mutation, R47L in two families; a novel missense mutation, R110W; and a novel frameshifting deletion, I298fsX307 in four families. Haplotype analysis, using polymorphic microsatellites from chromosome 6p23-24, revealed that R47L and I298fsX307 mutations arose either as ancient founders, or recurrent de novo mutations. Functional studies including: subcellular localization studies, EMSAs and luciferase-reporter assays, were undertaken and these demonstrated that: the R39X mutant failed to localize to the nucleus; the R47L and R110W mutants both lost DNA-binding ability; and the I298fsX307 mutant had reduced transactivational ability. In order to gain further insights, we undertook 3D-modeling of the GCMB DNA-binding domain, which revealed that the R110 residue is likely important for the structural integrity of helix 2, which forms part of the GCMB/DNA binding interface. Thus, our results, which expand the spectrum of hypoparathyroidism-associated GCMB mutations, help elucidate the molecular mechanisms underlying DNA-binding and transactivation that are required for this parathyroid-specific transcription factor.

  Funded by: Medical Research Council: G9825289

  Human molecular genetics 2010;19;10;2028-38

  PUBMED: 20190276; DOI: 10.1093/hmg/ddq084

• Underlying genetic models of inheritance in established type 2 diabetes associations.

  Salanti G, Southam L, Altshuler D, Ardlie K, Barroso I, Boehnke M, Cornelis MC, Frayling TM, Grallert H, Grarup N, Groop L, Hansen T, Hattersley AT, Hu FB, Hveem K, Illig T, Kuusisto J, Laakso M, Langenberg C, Lyssenko V, McCarthy MI, Morris A, Morris AD, Palmer CN, Payne F, Platou CG, Scott LJ, Voight BF, Wareham NJ, Zeggini E and Ioannidis JP

  Clinical and Molecular Epidemiology Unit and Clinical Trials and Evidence-Based Medicine Unit, Department of Hygiene and Epidemiology, School of Medicine, University of Ioannina, Ioannina, Greece.

  For most associations of common single nucleotide polymorphisms (SNPs) with common diseases, the genetic model of inheritance is unknown. The authors extended and applied a Bayesian meta-analysis approach to data from 19 studies on 17 replicated associations with type 2 diabetes. For 13 SNPs, the data fitted very well to an additive model of inheritance for the diabetes risk allele; for 4 SNPs, the data were consistent with either an additive model or a dominant model; and for 2 SNPs, the data were consistent with an additive or recessive model. Results were robust to the use of different priors and after exclusion of data for which index SNPs had been examined indirectly through proxy markers. The Bayesian meta-analysis model yielded point estimates for the genetic effects that were very similar to those previously reported based on fixed- or random-effects models, but uncertainty about several of the effects was substantially larger. The authors also examined the extent of between-study heterogeneity in the genetic model and found generally small between-study deviation values for the genetic model parameter. Heterosis could not be excluded for 4 SNPs. Information on the genetic model of robustly replicated association signals derived from genome-wide association studies may be useful for predictive modeling and for designing biologic and functional experiments.

  Funded by: NCRR NIH HHS: P20 RR021954-02, UL1 RR025752; Wellcome Trust: 077016, 079557, 088885, WT088885/Z/09/Z

  American journal of epidemiology 2009;170;5;537-45

  PUBMED: 19602701; PMC: 2732984; DOI: 10.1093/aje/kwp145

• Is the thrifty genotype hypothesis supported by evidence based on confirmed type 2 diabetes- and obesity-susceptibility variants?

  Southam L, Soranzo N, Montgomery SB, Frayling TM, McCarthy MI, Barroso I and Zeggini E

  Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK.

  According to the thrifty genotype hypothesis, the high prevalence of type 2 diabetes and obesity is a consequence of genetic variants that have undergone positive selection during historical periods of erratic food supply. The recent expansion in the number of validated type 2 diabetes- and obesity-susceptibility loci, coupled with access to empirical data, enables us to look for evidence in support (or otherwise) of the thrifty genotype hypothesis using proven loci.

  Methods: We employed a range of tests to obtain complementary views of the evidence for selection: we determined whether the risk allele at associated 'index' single-nucleotide polymorphisms is derived or ancestral, calculated the integrated haplotype score (iHS) and assessed the population differentiation statistic fixation index (F (ST)) for 17 type 2 diabetes and 13 obesity loci.

  Results: We found no evidence for significant differences for the derived/ancestral allele test. None of the studied loci showed strong evidence for selection based on the iHS score. We find a high F (ST) for rs7901695 at TCF7L2, the largest type 2 diabetes effect size found to date.

  Our results provide some evidence for selection at specific loci, but there are no consistent patterns of selection that provide conclusive confirmation of the thrifty genotype hypothesis. Discovery of more signals and more causal variants for type 2 diabetes and obesity is likely to allow more detailed examination of these issues.

  Funded by: Medical Research Council: G0601261, G0601261(80227); Wellcome Trust: 077016, 079557, 088885, WT077016/Z/05/Z, WT088885/Z/09/Z

  Diabetologia 2009;52;9;1846-51

  PUBMED: 19526209; PMC: 2723682; DOI: 10.1007/s00125-009-1419-3

• Functional analysis of the osteoarthritis susceptibility-associated GDF5 regulatory polymorphism.

  Egli RJ, Southam L, Wilkins JM, Lorenzen I, Pombo-Suarez M, Gonzalez A, Carr A, Chapman K and Loughlin J

  RMS Foundation, Bettlach, Switzerland.

  Objective: Single-nucleotide polymorphism (SNP) rs143383 (T to C) in the 5'-untranslated region (5'-UTR) of GDF5 has recently been reported to be associated with osteoarthritis (OA) susceptibility, with lower expression of the risk-associated T allele observed in vitro and in vivo. The in vivo studies were performed on cartilage tissue from OA patients. The present study was undertaken to expand the analysis of the effect of this SNP on GDF5 allelic expression to more joint tissue types, to investigate for cis and trans factors that interact with the SNP, and to examine novel cis-acting GDF5 regulatory polymorphisms.

  Methods: Tissue samples were collected from OA patients undergoing joint replacement of the hip or knee. Nucleic acid was extracted, and, using rs143383 and an assay that discriminates and quantifies allelic expression, the relative amount of GDF5 expression from the T and C alleles was measured. Additional common variants in the GDF5 transcript sequence were interrogated as potential regulatory elements using allelic expression and luciferase reporter assays, and electrophoretic mobility shift assays were used to search for trans factors binding to rs143383.

  Results: We observed a consistent allelic expression imbalance of GDF5 in all tissues tested, implying that the functional effect mediated by rs143383 on GDF5 expression is joint-wide. We identified a second polymorphism, located in the 3'-UTR of GDF5, that influenced allelic expression of the gene independent of rs143383. Finally, we observed differential binding of deformed epidermal autoregulatory factor 1 (DEAF-1) to the 2 alleles of rs143383.

  Conclusion: These findings show that the OA susceptibility mediated by polymorphism in GDF5 is not restricted to cartilage, emphasizing the need to consider the disease as involving the whole joint. The existence of an additional cis-acting regulatory polymorphism highlights the complexity of the regulation of expression of this important OA susceptibility locus. DEAF-1 is a trans-acting factor that merits further investigation as a potential tool for modulating GDF5 expression.

  Funded by: Arthritis Research UK

  Arthritis and rheumatism 2009;60;7;2055-64

  PUBMED: 19565498; DOI: 10.1002/art.24616

• Association of a functional microsatellite within intron 1 of the BMP5 gene with susceptibility to osteoarthritis.

  Wilkins JM, Southam L, Mustafa Z, Chapman K and Loughlin J

  University of Oxford, Institute of Musculoskeletal Sciences, Botnar Research Centre, Nuffield Orthopaedic Centre, Oxford, OX3 7LD, UK. james_wilkins@hms.harvard.edu

  Background: In a previous study carried out by our group, the genotyping of 36 microsatellite markers from within a narrow interval of chromosome 6p12.3-q13 generated evidence for linkage and for association to female hip osteoarthritis (OA), with the most compelling association found for a marker within intron 1 of the bone morphogenetic protein 5 gene (BMP5). In this study, we aimed to further categorize the association of variants within intron 1 of BMP5 with OA through an expanded genetic association study of the intron and subsequent functional analysis of associated polymorphisms.

  Methods: We genotyped 18 common polymorphisms including 8 microsatellites and 9 single nucleotide polymorphisms (SNPs) and 1 insertion/deletion (INDEL) from within highly conserved regions between human and mouse within intron 1 of BMP5. These markers were then tested for association to OA by a two-stage approach in which the polymorphisms were initially genotyped in a case-control cohort comprising 361 individuals with associated polymorphisms (P < or = 0.05) then genotyped in a second case-control cohort comprising 1185 individuals.

  Results: Two BMP5 intron 1 polymorphisms demonstrated association in the combined case-control cohort of 1546 individuals (765 cases and 781 controls): microsatellite D6S1276 (P = 0.018) and SNP rs921126 (P = 0.013). Functional analyses in osteoblastic, chondrocytic, and adipocytic cell lines indicated that allelic variants of D6S1276 have significant effects on the transcriptional activity of the BMP5 promoter in vitro.

  Conclusion: Variability in gene expression of BMP5 may be an important contributor to OA genetic susceptibility.

  Funded by: Arthritis Research UK: 16239

  BMC medical genetics 2009;10;141

  PUBMED: 20021689; PMC: 2807860; DOI: 10.1186/1471-2350-10-141

• Genome-wide association scan identifies a prostaglandin-endoperoxide synthase 2 variant involved in risk of knee osteoarthritis.

  Valdes AM, Loughlin J, Timms KM, van Meurs JJ, Southam L, Wilson SG, Doherty S, Lories RJ, Luyten FP, Gutin A, Abkevich V, Ge D, Hofman A, Uitterlinden AG, Hart DJ, Zhang F, Zhai G, Egli RJ, Doherty M, Lanchbury J and Spector TD

  Twin Research Unit, St. Thomas' Hospital Campus, Kings College London School of Medicine, London SE1 7EH, UK. ana.valdes@kcl.ac.uk

  Osteoarthritis (OA), the most prevalent form of arthritis in the elderly, is characterized by the degradation of articular cartilage and has a strong genetic component. Our aim was to identify genetic variants involved in risk of knee OA in women. A pooled genome-wide association scan with the Illumina550 Duo array was performed in 255 controls and 387 cases. Twenty-eight variants with p < 1 x 10(-5) were estimated to have probabilities of being false positives <or=0.5 and were genotyped individually in the original samples and in replication cohorts from the UK and the U.S. (599 and 272 cases, 1530 and 258 controls, respectively). The top seven associations were subsequently tested in samples from the Netherlands (306 cases and 584 controls). rs4140564 on chromosome 1 mapping 5' to both the PTGS2 and PLA2G4A genes was associated with risk of knee OA in all the cohorts studied (overall odds ratio OR(mh) = 1.55 95% C.I. 1.30-1.85, p < 6.9 x 10(-7)). Differential allelic expression analysis of PTGS2 with mRNA extracted from the cartilage of joint-replacement surgery OA patients revealed a significant difference in allelic expression (p < 1.0 x 10(-6)). These results suggest the existence of cis-acting regulatory polymorphisms that are in, or near to, PTGS2 and in modest linkage disequilibrium with rs4140564. Our results and previous studies on the role of the cyclooxygenase 2 enzyme encoded by PTGS2 underscore the importance of this signaling pathway in the pathogenesis of knee OA.

  Funded by: Wellcome Trust

  American journal of human genetics 2008;82;6;1231-40

  PUBMED: 18471798; PMC: 2427208; DOI: 10.1016/j.ajhg.2008.04.006

• A meta-analysis of European and Asian cohorts reveals a global role of a functional SNP in the 5' UTR of GDF5 with osteoarthritis susceptibility.

  Chapman K, Takahashi A, Meulenbelt I, Watson C, Rodriguez-Lopez J, Egli R, Tsezou A, Malizos KN, Kloppenburg M, Shi D, Southam L, van der Breggen R, Donn R, Qin J, Doherty M, Slagboom PE, Wallis G, Kamatani N, Jiang Q, Gonzalez A, Loughlin J and Ikegawa S

  Institute of Musculoskeletal Sciences, Botnar Research Centre, Nuffield Orthopaedic Centre, University of Oxford, Oxford, UK.

  We have performed a meta-analysis combining data for more than 11,000 individuals. It provides compelling evidence for a positive association between a functional single-nucleotide polymorphism (SNP) in the 5'-UTR of GDF5 (+104T/C; rs143383) and osteoarthritis (OA) in European and Asian populations. This SNP has recently been reported to be associated with OA in Japanese and Han Chinese populations. Attempts to replicate this association in European samples have been inconclusive, as no association was found in the case-control cohorts from the UK, Spain and Greece when studied individually. However, the pooled data of UK and Spain found an association of the T-allele with an odds ratio (OR) of 1.10. Although the European studies had adequate power to replicate the original findings from the Japanese cohort (OR = 1.79), these results suggest that the role of the GDF5 polymorphism may not be as strong in Europeans. To clarify whether the European studies were hampered by insufficient power, we combined new data from the UK and the Netherlands with the three published studies of Europe and Asia. The results provide strong evidence of a positive association of the GDF5 SNP with knee OA for Europeans as well as for Asians. The combined association for both ethnic groups is highly significant for the allele frequency model (P = 0.0004, OR = 1.21) and the dominant model (P < 0.0001, OR = 1.48). These findings represent the first highly significant evidence for a risk factor for the development of OA which affects two highly diverse ethnic groups.

  Human molecular genetics 2008;17;10;1497-504

  PUBMED: 18299287; DOI: 10.1093/hmg/ddn038

Ioanna Tachmazidou

it3@sanger.ac.uk Statistical Geneticist

I graduated from the Aristotle University of Thessaloniki in Greece in 2002 with a BSc in Mathematics and completed an MSc in Statistics at University College London in 2003. I then undertook a 4-years PhD in Bioinformatics at Imperial College London, in the first year of which I completed an MSc in Bioinformatics. I earned my PhD in Statistical Genetics under the supervision of Dr Maria De Iorio in 2008. Subsequently, I took up a Career Development Fellowship at the Biostatistics Unit of the MRC in Cambridge. I joined the Applied Statistical Genetics group as a staff scientist in 2011.

Research

My research interests are on all aspects of statistical genetics, and in particular in the genetic etiology of common disease. My current work is primarily focused on developing statistical methodology and software for the discovery of disease susceptibility loci. My research interests include: fine-scale mapping, design and analysis of large scale association studies, next generation sequence data and rare variants analysis, multivariate analysis of sequence data, Bayesian survival analysis, Bayesian inference and model selection.

I am involved in a number of international consortia, including the 500 exomes, UK10K, HELIC, and the INCHARGE projects.

References

• Bayesian semiparametric meta-analysis for genetic association studies.

  De Iorio M, Newcombe PJ, Tachmazidou I, Verzilli CJ and Whittaker JC

  Department of Epidemiology and Biostatistics, Imperial College, London, United Kingdom. m.deiorio@imperial.ac.uk

  We present a Bayesian semiparametric model for the meta-analysis of candidate gene studies with a binary outcome. Such studies often report results from association tests for different, possibly study-specific and non-overlapping genetic markers in the same genetic region. Meta-analyses of the results at each marker in isolation are seldom appropriate as they ignore the correlation that may exist between markers due to linkage disequilibrium (LD) and cannot assess the relative importance of variants at each marker. Also such marker-wise meta-analyses are restricted to only those studies that have typed the marker in question, with a potential loss of power. A better strategy is one which incorporates information about the LD between markers so that any combined estimate of the effect of each variant is corrected for the effect of other variants, as in multiple regression. Here we develop a Bayesian semiparametric model which models the observed genotype group frequencies conditional to the case/control status and uses pairwise LD measurements between markers as prior information to make posterior inference on adjusted effects. The approach allows borrowing of strength across studies and across markers. The analysis is based on a mixture of Dirichlet processes model as the underlying semiparametric model. Full posterior inference is performed through Markov chain Monte Carlo algorithms. The approach is demonstrated on simulated and real data.

  Genetic epidemiology 2011;35;5;333-40

  PUBMED: 21400586; DOI: 10.1002/gepi.20581

• Application of the optimal discovery procedure to genetic case-control studies: comparison with p values and asymptotic Bayes factors.

  Tachmazidou I, De Iorio M and Dudbridge F

  Medical Research Council, Biostatistics Unit, Institute of Public Health, Cambridge, UK. ioanna.tachmazidou@mrc-bsu.cam.ac.uk

  Objectives: The power of genetic association studies is limited by stringent levels of statistical significance. To improve power, Bayes factors (BFs) have been suggested as an alternative measure to the p value, and Storey recently introduced an optimal discovery procedure (ODP) for multiple testing. We aimed to adapt the ODP to genetic case-control studies and to compare its power to p values and asymptotic BFs (ABFs).

  Methods: We propose estimators of the ODP based on prospective and retrospective likelihoods. We performed simulations based on independent common SNPs and on sequence data including rare variants. Effects of causal SNPs were simulated under various distributions of effect size.

  Results: The true ODP is never outperformed, but the estimated ODP has similar power to p values and ABFs. For common SNPs the ODP offers power advantages only in extreme scenarios. However, for rare variants the ODP and ABF detect more associations at low false-positive rates than do p values.

  Conclusions: The ODP can provide higher power than p values for genetic case-control studies of common variants. However, as the ABF has similar power to the ODP and is computed more rapidly, it is our currently preferred method.

  Funded by: Medical Research Council: WBSU.1052.00.012.00001.01

  Human heredity 2011;71;1;37-49

  PUBMED: 21389730; DOI: 10.1159/000323518

• Bayesian variable selection for survival regression in genetics.

  Tachmazidou I, Johnson MR and De Iorio M

  Medical Research Council, Biostatistics Unit, Cambridge, United Kingdom. ioanna.tachmazidou@mrc-bsu.cam.ac.uk

  Variable selection in regression with very big numbers of variables is challenging both in terms of model specification and computation. We focus on genetic studies in the field of survival, and we present a Bayesian-inspired penalized maximum likelihood approach appropriate for high-dimensional problems. In particular, we employ a simple, efficient algorithm that seeks maximum a posteriori (MAP) estimates of regression coefficients. The latter are assigned a Laplace prior with a sharp mode at zero, and non-zero posterior mode estimates correspond to significant single nucleotide polymorphisms (SNPs). Using the Laplace prior reflects a prior belief that only a small proportion of the SNPs significantly influence the response. The method is fast and can handle datasets arising from imputation or resequencing. We demonstrate the localization performance, power and false-positive rates of our method in large simulation studies of dense-SNP datasets and sequence data, and we compare the performance of our method to the univariate Cox regression and to a recently proposed stochastic search approach. In general, we find that our approach improves localization and power slightly, while the biggest advantage is in false-positive counts and computing times. We also apply our method to a real prospective study, and we observe potential association between candidate ABC transporter genes and epilepsy treatment outcomes.

  Funded by: Wellcome Trust

  Genetic epidemiology 2010;34;7;689-701

  PUBMED: 20976796; DOI: 10.1002/gepi.20530

• Bayesian survival analysis in genetic association studies.

  Tachmazidou I, Andrew T, Verzilli CJ, Johnson MR and De Iorio M

  Department of Epidemiology and Public Health, Imperial College, London, UK. ioanna.tachmazidou@imperial.ac.uk

  Motivation: Large-scale genetic association studies are carried out with the hope of discovering single nucleotide polymorphisms involved in the etiology of complex diseases. There are several existing methods in the literature for performing this kind of analysis for case-control studies, but less work has been done for prospective cohort studies. We present a Bayesian method for linking markers to censored survival outcome by clustering haplotypes using gene trees. Coalescent-based approaches are promising for LD mapping, as the coalescent offers a good approximation to the evolutionary history of mutations.

  Results: We compare the performance of the proposed method in simulation studies to the univariate Cox regression and to dimension reduction methods, and we observe that it performs similarly in localizing the causal site, while offering a clear advantage in terms of false positive associations. Moreover, it offers computational advantages. Applying our method to a real prospective study, we observe potential association between candidate ABC transporter genes and epilepsy treatment outcomes.

  Availability: R codes are available upon request.

  Supplementary data are available at Bioinformatics online.

  Funded by: Wellcome Trust

  Bioinformatics (Oxford, England) 2008;24;18;2030-6

  PUBMED: 18617538; PMC: 2530885; DOI: 10.1093/bioinformatics/btn351

• Genetic association mapping via evolution-based clustering of haplotypes.

  Tachmazidou I, Verzilli CJ and De Iorio M

  Department of Epidemiology and Public Health, Imperial College London, United Kingdom. ioanna.tachmazidou03@ic.ac.uk

  Multilocus analysis of single nucleotide polymorphism haplotypes is a promising approach to dissecting the genetic basis of complex diseases. We propose a coalescent-based model for association mapping that potentially increases the power to detect disease-susceptibility variants in genetic association studies. The approach uses Bayesian partition modelling to cluster haplotypes with similar disease risks by exploiting evolutionary information. We focus on candidate gene regions with densely spaced markers and model chromosomal segments in high linkage disequilibrium therein assuming a perfect phylogeny. To make this assumption more realistic, we split the chromosomal region of interest into sub-regions or windows of high linkage disequilibrium. The haplotype space is then partitioned into disjoint clusters, within which the phenotype-haplotype association is assumed to be the same. For example, in case-control studies, we expect chromosomal segments bearing the causal variant on a common ancestral background to be more frequent among cases than controls, giving rise to two separate haplotype clusters. The novelty of our approach arises from the fact that the distance used for clustering haplotypes has an evolutionary interpretation, as haplotypes are clustered according to the time to their most recent common ancestor. Our approach is fully Bayesian and we develop a Markov Chain Monte Carlo algorithm to sample efficiently over the space of possible partitions. We compare the proposed approach to both single-marker analyses and recently proposed multi-marker methods and show that the Bayesian partition modelling performs similarly in localizing the causal allele while yielding lower false-positive rates. Also, the method is computationally quicker than other multi-marker approaches. We present an application to real genotype data from the CYP2D6 gene region, which has a confirmed role in drug metabolism, where we succeed in mapping the location of the susceptibility variant within a small error.

  Funded by: Wellcome Trust

  PLoS genetics 2007;3;7;e111

  PUBMED: 17616979; PMC: 1913101; DOI: 10.1371/journal.pgen.0030111