Heather J Cordell

Newcastle University, Newcastle-on-Tyne, England, United Kingdom

Are you Heather J Cordell?

Claim your profile

Publications (171)1250.99 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Atopic dermatitis and psoriasis are the two most common immune-mediated inflammatory disorders affecting the skin. Genome-wide studies demonstrate a high degree of genetic overlap, but these diseases have mutually exclusive clinical phenotypes and opposing immune mechanisms. Despite their prevalence, atopic dermatitis and psoriasis very rarely co-occur within one individual. By utilizing genome-wide association study and ImmunoChip data from >19,000 individuals and methodologies developed from meta-analysis, we have identified opposing risk alleles at shared loci as well as independent disease-specific loci within the epidermal differentiation complex (chromosome 1q21.3), the Th2 locus control region (chromosome 5q31.1), and the major histocompatibility complex (chromosome 6p21-22). We further identified previously unreported pleiotropic alleles with opposing effects on atopic dermatitis and psoriasis risk in PRKRA and ANXA6/TNIP1. In contrast, there was no evidence for shared loci with effects operating in the same direction on both diseases. Our results show that atopic dermatitis and psoriasis have distinct genetic mechanisms with opposing effects in shared pathways influencing epidermal differentiation and immune response. The statistical analysis methods developed in the conduct of this study have produced additional insight from previously published data sets. The approach is likely to be applicable to the investigation of the genetic basis of other complex traits with overlapping and distinct clinical features. Copyright © 2015 The Authors. Published by Elsevier Inc. All rights reserved.
    The American Journal of Human Genetics 01/2015; 96(1):104-120. · 11.20 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Inherited abnormalities of complement are found in ∼60% of patients with atypical haemolytic uraemic syndrome (aHUS). Such abnormalities are not fully penetrant. In this study, we have estimated the penetrance of the disease in three families with a CFH mutation (c.3643C>G; p. Arg1215Gly) in whom a common lineage is probable. 25 individuals have been affected with aHUS with three peaks of incidence-early childhood (n=6), early adulthood (n=11) and late adulthood (n=8). Eighteen individuals who have not developed aHUS carry the mutation.
    Journal of Medical Genetics 09/2014; · 5.64 Impact Factor
  • Source
    Heather J Cordell
    [Show abstract] [Hide abstract]
    ABSTRACT: I present a summary of the results and discussions held within the working group on gene-based tests at Genetic Analysis Workshop 18 (GAW18). The main focus of interest in our working group was modeling the action of combinations or "groups" of genetic variants, with a group of variants most often defined as a set of single-nucleotide polymorphisms lying within a known gene. Some contributions investigated the performance of previously proposed methods (particularly rare variant collapsing or burden-type methods) for addressing this question, applied to the GAW18 data, and other contributions developed novel approaches and addressed novel questions. Most approaches were successful in detecting significant effects at MAP4 in the simulated data. No other genetic effects were consistently detected across different analyses. Low power was noted, particularly for those methods that restricted analysis to purely the subset of unrelated individuals.
    Genetic Epidemiology 09/2014; 38 Suppl 1(S1):S44-8. · 2.95 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Urofacial syndrome (UFS) is an autosomal recessive congenital disease featuring grimacing and incomplete bladder emptying. Mutations of HPSE2, encoding heparanase 2, a heparanase 1 inhibitor, occur in UFS, but knowledge about the HPSE2 mutation spectrum is limited. Here, seven UFS kindreds with HPSE2 mutations are presented, including one with deleted asparagine 254, suggesting a role for this amino acid, which is conserved in vertebrate orthologs. HPSE2 mutations were absent in 23 non-neurogenic neurogenic bladder probands and, of 439 families with nonsyndromic vesicoureteric reflux, only one carried a putative pathogenic HPSE2 variant. Homozygous Hpse2 mutant mouse bladders contained urine more often than did wild-type organs, phenocopying human UFS. Pelvic ganglia neural cell bodies contained heparanase 1, heparanase 2, and leucine-rich repeats and immunoglobulin-like domains-2 (LRIG2), which is mutated in certain UFS families. In conclusion, heparanase 2 is an autonomic neural protein implicated in bladder emptying, but HPSE2 variants are uncommon in urinary diseases resembling UFS
    Journal of the American Society of Nephrology 08/2014; · 9.47 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Approaches based on linear mixed models (LMMs) have recently gained popularity for modelling population substructure and relatedness in genome-wide association studies. In the last few years, a bewildering variety of different LMM methods/software packages have been developed, but it is not always clear how (or indeed whether) any newly-proposed method differs from previously-proposed implementations. Here we compare the performance of several LMM approaches (and software implementations, including EMMAX, GenABEL, FaST-LMM, Mendel, GEMMA and MMM) via their application to a genome-wide association study of visceral leishmaniasis in 348 Brazilian families comprising 3626 individuals (1972 genotyped). The implementations differ in precise details of methodology implemented and through various user-chosen options such as the method and number of SNPs used to estimate the kinship (relatedness) matrix. We investigate sensitivity to these choices and the success (or otherwise) of the approaches in controlling the overall genome-wide error-rate for both real and simulated phenotypes. We compare the LMM results to those obtained using traditional family-based association tests (based on transmission of alleles within pedigrees) and to alternative approaches implemented in the software packages MQLS, ROADTRIPS and MASTOR. We find strong concordance between the results from different LMM approaches, and all are successful in controlling the genome-wide error rate (except for some approaches when applied naively to longitudinal data with many repeated measures). We also find high correlation between LMMs and alternative approaches (apart from transmission-based approaches when applied to SNPs with small or non-existent effects). We conclude that LMM approaches perform well in comparison to competing approaches. Given their strong concordance, in most applications, the choice of precise LMM implementation cannot be based on power/type I error considerations but must instead be based on considerations such as speed and ease-of-use.
    PLoS Genetics 07/2014; 10(7):e1004445. · 8.52 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Our understanding of genome biology, genomics, and disease, and even hu-man history, has advanced tremen-dously with the completion of the Human Genome Project. Technologi-cal advances coupled with significant cost reductions in genomic research have yielded novel insights into disease etiol-ogy, diagnosis, and therapy for some of the world's most intractable and devastat-ing diseases—including ma-laria, HIV/AIDS, tuberculosis, cancer, and diabetes. Yet, de-spite the burden of infectious diseases and, more recently, noncommunicable diseases (NCDs) in Africa, Africans have only par-ticipated minimally in genomics research. Of the thousands of genome-wide association studies (GWASs) that have been conducted globally, only seven (for HIV susceptibility, malaria, tuberculosis, and podoconiosis) have been conducted exclusively on Afri-can participants; four others (for prostate cancer, obsessive compulsive disorder, and anthropometry) included some African participants (www.genome.gov/gwastudies/). As discussed in 2011 (www.h3africa.org), if the dearth of genomics research involving Africans persists, the potential health and economic benefits emanating from genomic science may elude an entire continent. The lack of large-scale genomics studies in Africa is the result of many deep-seated issues, including a shortage of African scien-tists with genomic research expertise, lack of biomedical research infrastructure, lim-ited computational expertise and resources, lack of adequate support for biomedical research by African governments, and the participation of many African scientists in collaborative research at no more than the level of sample collection. Overcoming these limitations will, in part, depend on African Enabling the genomic revolution in Africa By The H3Africa Consortium * H3Africa is developing capacity for health-related genomics research in Africa Yet, roughly a decade ago, newly pro-posed DNA-based taxonomy (11) promised to solve the species debate. A Barcode of Life Data Systems (BOLD) (12) quickly emerged, seeking to provide a reliable, cost-effective solution to the problem of species identification (12) and a standard screening threshold of sequence differ-ence (10× average intraspecific difference) to speed the discovery of new animal spe-cies (13). Sometimes considered a "carica-ture of real taxonomy" (14), this approach failed to identify, perhaps not surprisingly, two American crow species and a number of members of the herring gull Larus ar-gentatus species assemblage above the set threshold (13). Furthermore, despite past (3) and present (6) sequencing projects, carrion crows and hooded crows can also not be differentiated from one another by means of DNA-barcode approaches. By contrast, Poelstra et al. show that much more DNA sequencing data are needed, combined with RNA expression data, to reconstruct the evolution of a reproductive barrier that culminated in the speciation of these two crow taxa. Armed with this new very detailed genetic informa-tion, it is clear that none of the currently formulated species concepts fully apply to these two crow taxa (unless one is willing relax some stringency in the various definitions). In-deed, the genomes of German carrion crows are much more similar to those of hooded crows than to Spanish car-rion crows. Put simply, apart from the few carrion crow type "speciation islands," German carrion crows could be con-sidered to represent hooded crows with a black (carrion crow) phenotype. There is a clear need for ad-ditional population genomic studies using a more dense sampling, especially among the fully black carrion crows, before the complexity of repro-ductive isolation and speciation among these two taxa can be fully understood. The specia-tion genomics strategy already proved itself in unraveling the complexities of mimicry among many Heliconius butterfly taxa (7) and, as in the study of Poelstra et al., stresses the im-portance of using RNA-based information in addition to DNA. Only time will tell if, and when, German carrion crows will adopt the "hooded phenotype," a fate that seems un-avoidable. Until then, we can only applaud these crows for defeating Linnaeus's curse.
    Science 06/2014; 344(6190):1346-1348. · 31.48 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Genetic Analysis Workshop 18 provided a platform for developing and evaluating statistical methods to analyze whole-genome sequence data from a pedigree-based sample. In this article we present an overview of the data sets and the contributions that analyzed these data. The family data, donated by the Type 2 Diabetes Genetic Exploration by Next-Generation Sequencing in Ethnic Samples Consortium, included sequence-level genotypes based on sequencing and imputation, genome-wide association genotypes from prior genotyping arrays, and phenotypes from longitudinal assessments. The contributions from individual research groups were extensively discussed before, during, and after the workshop in theme-based discussion groups before being submitted for publication.
    BMC proceedings 06/2014; 8(1):S1.
  • Source
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The genetic contribution to the variation in human lifespan is approximately 25%. Despite the large number of identified disease-susceptibility loci, it is not known which loci influence population mortality.We performed a genome-wide association meta-analysis of 7729 long-lived individuals of European descent (≥ 85 years) and 16121 younger controls (< 65 years) followed by replication in an additional set of 13060 long-lived individuals and 61156 controls. In addition, we performed a subset analysis in cases≥90 years.We observed genome-wide significant association with longevity, as reflected by survival to ages beyond 90 years, at a novel locus, rs2149954, on chromosome 5q33.3 (OR=1.10, P =1.74 x 10(-8)). We also confirmed association of rs4420638 on chromosome 19q13.32 (OR=0.72, P=3.40 x 10(-36)), representing the TOMM40/APOE/APOC1 locus. In a prospective meta-analysis (n=34103) the minor allele of rs2149954 (T) on chromosome 5q33.3 associates with increased survival (HR=0.95, P=0.003). This allele has previously been reported to associate with low blood pressure in middle age. Interestingly, the minor allele (T) associates with decreased cardiovascular mortality risk, independent of blood pressure.We report on the first GWAS-identified longevity locus on chromosome 5q33.3 influencing survival in the general European population. The minor allele of this locus associates with low blood pressure in middle age, although the contribution of this allele to survival may be less dependent on blood pressure. Hence, the pleiotropic mechanisms by which this intragenic variation contributes to lifespan regulation have to be elucidated.
    Human Molecular Genetics 03/2014; · 6.68 Impact Factor
  • Source
    Richard Howey, Heather J Cordell
    [Show abstract] [Hide abstract]
    ABSTRACT: Genome-wide association studies allow detection of non-genotyped disease-causing variants through testing of nearby genotyped SNPs. This approach may fail when there are no genotyped SNPs in strong LD with the causal variant. Several genotyped SNPs in weak LD with the causal variant may, however, considered together, provide equivalent information. This observation motivates popular but computationally intensive approaches based on imputation or haplotyping. Here we present a new method and accompanying software designed for this scenario. Our approach proceeds by selecting, for each genotyped "anchor" SNP, a nearby genotyped "partner" SNP, chosen via a specific algorithm we have developed. These two SNPs are used as predictors in linear or logistic regression analysis to generate a final significance test. In simulations, our method captures much of the signal captured by imputation, while taking a fraction of the time and disc space, and generating a smaller number of false-positives. We apply our method to a case/control study of severe malaria genotyped using the Affymetrix 500K array. Previous analysis showed that fine-scale sequencing of a Gambian reference panel in the region of the known causal locus, followed by imputation, increased the signal of association to genome-wide significance levels. Our method also increases the signal of association from P≈2×10-6 to P≈6×10-11. Our method thus, in some cases, eliminates the need for more complex methods such as sequencing and imputation, and provides a useful additional test that may be used to identify genetic regions of interest.
    Genetic Epidemiology 02/2014; · 2.95 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Gene variants known to contribute to Autoimmune Addison's disease (AAD) susceptibility include those at the MHC, MICA, CIITA, CTLA4, PTPN22, CYP27B1, NLRP-1 and CD274 loci. The majority of the genetic component to disease susceptibility has yet to be accounted for. To investigate the role of 19 candidate genes in AAD susceptibility in six European case-control cohorts. A sequential association study design was employed with genotyping using Sequenom iPlex technology. In phase one, 85 SNPs in 19 genes were genotyped in UK and Norwegian AAD cohorts (691 AAD, 715 controls). In phase two, 21 SNPs in 11 genes were genotyped in German, Swedish, Italian and Polish cohorts (1264 AAD, 1221 controls). In phase three, to explore association of GATA3 polymorphisms with AAD and to determine if this association extended to other autoimmune conditions, 15 SNPs in GATA3 were studied in UK and Norwegian AAD cohorts, 1195 type 1 diabetes patients from Norway, 650 rheumatoid arthritis patients from New Zealand and in 283 UK Graves' disease patients. Meta-analysis was used to compare genotype frequencies between the participating centres, allowing for heterogeneity. We report significant association with alleles of two STAT4 markers in AAD cohorts (rs4274624: P = 0.00016; rs10931481: P = 0.0007). In addition, nominal association of AAD with alleles at GATA3 was found in 3 patient cohorts and supported by meta-analysis. Association of AAD with CYP27B1 alleles was also confirmed, which replicates previous published data. Finally, nominal association was found at SNPs in both the NF-κB1 and IL23A genes in the UK and Italian cohorts respectively. Variants in the STAT4 gene, previously associated with other autoimmune conditions, confer susceptibility to AAD. Additionally, we report association of GATA3 variants with AAD: this adds to the recent report of association of GATA3 variants with rheumatoid arthritis.
    PLoS ONE 01/2014; 9(3):e88991. · 3.53 Impact Factor
  • Source
    Kristin L Ayers, Heather J Cordell
    [Show abstract] [Hide abstract]
    ABSTRACT: Under the premise that multiple causal variants exist within a disease gene and that we are underpowered to detect these variants individually, a variety of methods have been developed that attempt to cluster rare variants within a gene so that the variants may gather strength from one another. These methods group variants by gene or proximity, and test one gene or marker window at a time. We propose analyzing all genes simultaneously with a penalized regression method that enables grouping of all (rare and common) variants within a gene while subgrouping rare variants, thus borrowing strength from both rare and common variants within the same gene. We apply this approach using a burden based weighting of the rare variants to the Genetic Analysis Workshop 18 data.
    BMC proceedings 01/2014; 8(Suppl 1):S43.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Primary vesicoureteric reflux (VUR), the retrograde flow of urine from the bladder toward the kidneys, results from a developmental anomaly of the vesicoureteric valve mechanism, and is often associated with other urinary tract anomalies. It is the most common urological problem in children, with an estimated prevalence of 1-2%, and is a major cause of hypertension in childhood and of renal failure in childhood or adult life. We present the results of a genetic linkage and association scan using 900,000 markers. Our linkage results show a large number of suggestive linkage peaks, with different results in two groups of families, suggesting that VUR is even more genetically heterogeneous than previously imagined. The only marker achieving P < 0.02 for linkage in both groups of families is 270 kb from EMX2. In three sibships, we found recessive linkage to KHDRBS3, previously reported in a Somali family. In another family we discovered sex-reversal associated with VUR, implicating PRKX, for which there was weak support for dominant linkage in the overall data set. Several other candidate genes are suggested by our linkage or association results, and four of our linkage peaks are within copy-number variants recently found to be associated with renal hypodysplasia. Undoubtedly there are many genes related to VUR. Our study gives support to some loci suggested by earlier studies as well as suggesting new ones, and provides numerous indications for further investigations.
    Molecular genetics & genomic medicine. 01/2014; 2(1):7-29.
  • Source
    Science. 01/2014; 344(6190):1346-8.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In the last few years, a bewildering variety of methods/software packages that use linear mixed models to account for sample relatedness on the basis of genome-wide genomic information have been proposed. We compared these approaches as implemented in the programs EMMAX, FaST-LMM, Gemma, and GenABEL (FASTA/GRAMMAR-Gamma) on the Genetic Analysis Workshop 18 data. All methods performed quite similarly and were successful in reducing the genomic control inflation factor to reasonable levels, particularly when the mean values of the observations were used, although more variation was observed when data from each time point were used individually. From a practical point of view, we conclude that it makes little difference to the results which method/software package is used, and the user can make the choice of package on the basis of personal taste or computational speed/convenience.
    BMC proceedings 01/2014; 8(Suppl 1 Genetic Analysis Workshop 18Vanessa Olmo):S79.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Atopic dermatitis (AD) is a major inflammatory condition of the skin caused by inherited skin barrier deficiency, with mutations in the filaggrin gene predisposing to development of AD. Support for barrier deficiency initiating AD came from flaky tail mice, which have a frameshift mutation in Flg and also carry an unknown gene, matted, causing a matted hair phenotype. We sought to identify the matted mutant gene in mice and further define whether mutations in the human gene were associated with AD. A mouse genetics approach was used to separate the matted and Flg mutations to produce congenic single-mutant strains for genetic and immunologic analysis. Next-generation sequencing was used to identify the matted gene. Five independently recruited AD case collections were analyzed to define associations between single nucleotide polymorphisms (SNPs) in the human gene and AD. The matted phenotype in flaky tail mice is due to a mutation in the Tmem79/Matt gene, with no expression of the encoded protein mattrin in the skin of mutant mice. Matt(ft) mice spontaneously have dermatitis and atopy caused by a defective skin barrier, with mutant mice having systemic sensitization after cutaneous challenge with house dust mite allergens. Meta-analysis of 4,245 AD cases and 10,558 population-matched control subjects showed that a missense SNP, rs6694514, in the human MATT gene has a small but significant association with AD. In mice mutations in Matt cause a defective skin barrier and spontaneous dermatitis and atopy. A common SNP in MATT has an association with AD in human subjects.
    The Journal of allergy and clinical immunology 09/2013; · 12.05 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Twenty percent of people aged 20 to 79 have type 2 diabetes (T2D) in the United Arab Emirates (UAE). Genome-wide association studies (GWAS) to identify genes for T2D have not been reported for Arab countries. We performed a discovery GWAS in an extended UAE family (N = 178; 66 diabetic; 112 healthy) genotyped on the Illumina Human 660 Quad Beadchip, with independent replication of top hits in 116 cases and 199 controls. Power to achieve genome-wide significance (commonly P = 5 × 10(-8) ) was therefore limited. Nevertheless, transmission disequilibrium testing in FBAT identified top hits at Chromosome 4p12-p13 (KCTD8: rs4407541, P = 9.70 × 10(-6) ; GABRB1: rs10517178/rs1372491, P = 4.19 × 10(-6) ) and 14q13 (PRKD1: rs10144903, 3.92 × 10(-6) ), supported by analysis using a linear mixed model approximation in GenABEL (4p12-p13 GABRG1/GABRA2: rs7662743, Padj-agesex = 2.06 × 10(-5) ; KCTD8: rs4407541, Padj-agesex = 1.42 × 10(-4) ; GABRB1: rs10517178/rs1372491, Padj-agesex = 0.027; 14q13 PRKD1: rs10144903, Padj-agesex = 6.95 × 10(-5) ). SNPs across GABRG1/GABRA2 did not replicate, whereas more proximal SNPs rs7679715 (Padj-agesex = 0.030) and rs2055942 (Padj-agesex = 0.022) at COX7B2/GABRA4 did, in addition to a trend distally at KCTD8 (rs4695718: Padj-agesex = 0.096). Modelling of discovery and replication data support independent signals at GABRA4 (rs2055942: Padj-agesex-combined = 3 × 10(-4) ) and at KCTD8 (rs4695718: Padj-agesex-combined = 2 × 10(-4) ). Replication was observed for PRKD1 rs1953722 (proxy for rs10144903; Padj-agesex = 0.031; Padj-agesex-combined = 2 × 10(-4) ). These genes may provide important functional leads in understanding disease pathogenesis in this population.
    Annals of Human Genetics 08/2013; · 1.93 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: -Association between the C677T polymorphism of the methylene tetrahydrofolate reductase (MTHFR) gene and congenital heart disease (CHD) is contentious. -We compared genotypes between CHD cases and controls, and between mothers of CHD cases and controls. We placed our results in context by conducting meta-analyses of previously published studies. Among 5,814 cases with primary genotype data and 10,056 controls, there was no evidence of association between MTHFR C677T genotype and CHD risk (OR 0.96 [95% CI 0.87-1.07]). A random-effects meta-analysis of all studies (involving 7,697 cases and 13,125 controls) suggested the presence of association (OR 1.25 [95% CI 1.03-1.51]; p=0.022), but with substantial heterogeneity among contributing studies (I(2)=64.4%), and evidence of publication bias. Meta-analysis of large studies only (defined by a variance of the log OR less than 0.05), which together contributed 83% of all cases, yielded no evidence of association (OR 0.97 [95% CI 0.91-1.03]), without significant heterogeneity (I(2)=0). Moreover, meta-analysis of 1,781 mothers of CHD cases (829 of whom were genotyped in this study) and 19,861 controls revealed no evidence of association between maternal C677T genotype and risk of CHD in offspring (OR 1.13 [95% CI 0.87-1.47]). There was no significant association between MTHFR genotype and CHD risk in large studies from regions with different levels of dietary folate. -The MTHFR C677T polymorphism, which directly influences plasma folate levels, is not associated with CHD risk. Publication biases appear to substantially contaminate the literature with regard to this genetic association.
    Circulation Cardiovascular Genetics 07/2013; · 5.34 Impact Factor
  • Source
  • Source

Publication Stats

9k Citations
1,250.99 Total Impact Points


  • 2006–2015
    • Newcastle University
      • • Institute of Genetic Medicine
      • • School of Mathematics and Statistics
      Newcastle-on-Tyne, England, United Kingdom
    • National Human Genome Research Institute
      Maryland, United States
  • 2013
    • University of Newcastle
      Newcastle, New South Wales, Australia
    • Leiden University Medical Centre
      Leyden, South Holland, Netherlands
    • United Arab Emirates University
      Al Ain, Abu Dhabi, United Arab Emirates
  • 2012
    • Yamagata University
      Ямагата, Yamagata, Japan
  • 2008–2012
    • Wellcome Trust Sanger Institute
      Cambridge, England, United Kingdom
    • Newcastle University Medicine Malaysia
      Bharu, Johor, Malaysia
  • 2011
    • WWF United Kingdom
      Londinium, England, United Kingdom
  • 2000–2011
    • University of Cambridge
      • Department of Medical Genetics
      Cambridge, ENG, United Kingdom
  • 2009
    • University of Western Australia
      • Centre for Health Services Research
      Perth, Western Australia, Australia
  • 2008–2009
    • Institute of Human Genetics
      Amadavad, Gujarāt, India
  • 2000–2008
    • Cambridge Institute for Medical Research
      Cambridge, England, United Kingdom
  • 2007
    • Université Paris-Sud 11
      Orsay, Île-de-France, France
    • University of California, San Francisco
      • Department of Epidemiology and Biostatistics
      San Francisco, CA, United States
  • 2001
    • Providence University
      臺中市, Taiwan, Taiwan
    • Università degli studi di Cagliari
      • Department of Biomedical Science
      Cagliari, Sardinia, Italy
  • 1997–2001
    • Case Western Reserve University
      • • Department of Epidemiology and Biostatistics
      • • Rammelkamp Center for Education and Research
      Cleveland, OH, United States
  • 1993–1998
    • University of Oxford
      • Wellcome Trust Centre for Human Genetics
      Oxford, ENG, United Kingdom