Pak C Sham

Lands Department of The Government of the Hong Kong Special Administrative Region, Hong Kong, Hong Kong

Are you Pak C Sham?

Claim your profile

Publications (546)3064.97 Total impact

  • Jia-En Deng · Pak C Sham · Miao-Xin Li ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Availability: SNPTracker is freely available for download at
    G3-Genes Genomes Genetics 11/2015; DOI:10.1534/g3.115.021832 · 3.20 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Objective: Multiple etiological and prognostic factors have been implied in schizophrenia and its outcome. Advanced paternal age has been reported as a risk factor in schizophrenia. Whether this may affect schizophrenia outcome was not previously studied. We hypothesized that advanced paternal age may have a negative effect on the outcome of relapse in schizophrenia. Method: We interviewed 191 patients with first-episode schizophrenia and their relatives for parental ages, sociodemographic factors at birth, birth rank, family history of psychotic disorders, and obstetric complications. The outcome measure was the presence of relapse at the end of the first year of treatment. Results: In the 1-year follow-up period, 42 (22%) patients experienced 1 or more relapses. The mean paternal age was 34.62 years (SD 7.69). Patients who relapsed had significantly higher paternal age, poorer medication adherence, were female, and were hospitalized at onset, compared with patients who did not relapse. A multivariate regression analysis showed that advanced paternal age (OR 1.05, 95% CI 1.01 to 1.10), medication nonadherence (OR 2.37, 95% CI 1.12 to 4.99), and female sex (OR 2.44, 95% CI 1.14 to 5.24) independently contributed to a higher risk of relapse. Analysis between different paternal age groups found a significantly higher relapse rate with paternal age over 40. Conclusions: Advanced paternal age is found to be modestly but significantly related to more relapses, and such an effect is the strongest at a cut-off of paternal age of 40 years or older. The effect is less likely to be mediated through less effective parental supervision or nonadherence to medication. Other possible biological mechanisms need further explorations.
    Canadian journal of psychiatry. Revue canadienne de psychiatrie 10/2015; 60(8):346-353. · 2.55 Impact Factor
  • Source
    Q Li · Y O Leung · I Zhou · L C Ho · W Kong · P Basil · R Wei · S Lam · X Zhang · A C K Law · S E Chua · P C Sham · E X Wu · G M McAlonan ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Prenatal exposure to maternal immune activation (MIA) increases the risk of schizophrenia and autism in the offspring. The MIA rodent model provides a valuable tool to directly test the postnatal consequences of exposure to an early inflammatory insult; and examine novel preventative strategies. Here we tested the hypotheses that behavioural differences in the MIA mouse model are accompanied by in vivo and ex vivo alterations in brain biochemistry; and that these can be prevented by a post-weaning diet enriched with n-3 polyunsaturated fatty acid (PUFA). The viral analogue PolyI:C (POL) or saline (SAL) was administered to pregnant mice on gestation day 9. Half the resulting male offspring (POL=21; SAL=17) were weaned onto a conventional lab diet (n-6 PUFA); half were weaned onto n-3 PUFA-enriched diet. In vivo magnetic resonance spectroscopy measures were acquired prior to behavioural tests; glutamic acid decarboxylase 67 (GAD67) and tyrosine hydroxylase protein levels were measured ex vivo. The main findings were: (i) Adult MIA-exposed mice fed a standard diet had greater N-acetylaspartate/creatine (Cr) and lower myo-inositol/Cr levels in the cingulate cortex in vivo. (ii) The extent of these metabolite differences was correlated with impairment in prepulse inhibition. (iii) MIA-exposed mice on the control diet also had higher levels of anxiety and altered levels of GAD67 ex vivo. (iv) An n-3 PUFA diet prevented all the in vivo and ex vivo effects of MIA observed. Thus, n-3 PUFA dietary enrichment from early life may offer a relatively safe and non-toxic approach to limit the otherwise persistent behavioural and biochemical consequences of prenatal exposure to inflammation. This result may have translational importance.
    Translational Psychiatry 09/2015; 5(9):e641. DOI:10.1038/tp.2015.126 · 5.62 Impact Factor
  • Johnny Sh Kwan · Miao-Xin Li · Jia-En Deng · Pak C Sham ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Imputing individual-level genotypes (or genotype imputation) is now a standard procedure in genome-wide association studies (GWAS) to examine disease associations at untyped common genetic variants. Meta-analysis of publicly available GWAS summary statistics can allow more disease-associated loci to be discovered, but these data are usually provided for various variant sets. Thus imputing these summary statistics of different variant sets into a common reference panel for meta-analyses is impossible using traditional genotype imputation methods. Here we develop a Fast and Accurate P-value Imputation (FAPI) method that utilizes summary statistics of common variants only. Its computational cost is linear with the number of untyped variants and has similar accuracy compared with IMPUTE2 with prephasing, one of the leading methods in genotype imputation. In addition, based on the FAPI idea, we develop a metric to detect abnormal association at a variant and showed that it had a significantly greater power compared with LD-PAC, a method that quantifies the evidence of spurious associations based on likelihood ratio. Our method is implemented in a user-friendly software tool, which is available at Journal of Human Genetics advance online publication, 26 August 2015; doi:10.1038/ejhg.2015.190.
    European journal of human genetics: EJHG 08/2015; DOI:10.1038/ejhg.2015.190 · 4.35 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Previous genome-wide association studies (GWAS), which were mainly based on single variant analysis, have identified many susceptibility loci for systemic lupus erythematosus (SLE). However, the genetic architecture of this complex disease is far from understood. A gene-based analysis may help identify novel loci by considering global evidence of association from a gene or a genomic region rather than focusing on evidence from individual variants. Based on meta-analysis results of two SLE GWASs, we performed gene- and region-based analysis followed by replication with a total of 4,626 cases and 7,466 controls of Asian ancestry. Allelic differential expression was measured by pyrosequencing. Over half of the reported SLE susceptibility loci showed evidence of independent effects, which is important for understanding the mechanisms of association and explaining disease heritability. ANXA6 was detected as a novel susceptibility gene for SLE with several SNPs contributing independently to disease association. The risk allele of rs11960458 correlated significantly with increased expression of ANXA6 in PBMCs of heterozygous healthy controls. Several other associated SNPs may also regulate ANXA6 expression based on data from public databases. Higher expression of ANXA6 in SLE cases was also reported previously. Our study demonstrated the merit of gene-based analysis in identifying novel susceptibility loci, especially those with independent effects, and also demonstrated a widespread presence of independent contributors in susceptibility genes for SLE. This article is protected by copyright. All rights reserved. © 2015, American College of Rheumatology.
    Arthritis and Rheumatology 07/2015; DOI:10.1002/art.39275
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Diffuse oesophageal leiomyomatosis (DOL) is a rare disorder characterized by tumorous overgrowth of the muscular wall of the oesophagus. DOL is present in 5 % of Alport syndrome (AS) patients. AS is a rare hereditary disease that involves varying degrees of hearing impairment, ocular changes and progressive glomerulonephritis leading to renal failure. In DOL-AS patients, the genetic defect consists of a deletion involving the COL4A5 and COL4A6 genes on the X chromosome. We report a two-generation family (4 individuals; parents and two children, one male and one female) with two members (mother and son) affected with oesophageal leiomyomatosis. Signs of potential renal failure, which characterizes AS, were only apparent in the index patient (son) 2 years and three months after the initial diagnosis of DOL. Blood DNA from the four family members were submitted to exome sequencing and array genotyping to perform a genome wide screening for disease causal single nucleotide (SN) and copy number (CN) variations. Analyses revealed a new 40kb deletion encompassing from intron 2 of COL4A5 to intron 1 of COL4A6 at Xq22.3. The breakpoints were also identified. Possible confounding pathogenic exonic variants in genes known to be involved in other extracellular matrices disorders were also shared by the two affected individuals. Meticulous analysis of the maternal DNA revealed a case of gonosomal mosaicism. This is the first report of gonadosomal mosaicism associated to DOL-AS.
    BMC Medical Genetics 07/2015; 16(1):49. DOI:10.1186/s12881-015-0189-7 · 2.08 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Over a quarter of drugs that enter clinical development fail because they are ineffective. Growing insight into genes that influence human disease may affect how drug targets and indications are selected. However, there is little guidance about how much weight should be given to genetic evidence in making these key decisions. To answer this question, we investigated how well the current archive of genetic evidence predicts drug mechanisms. We found that, among well-studied indications, the proportion of drug mechanisms with direct genetic support increases significantly across the drug development pipeline, from 2.0% at the preclinical stage to 8.2% among mechanisms for approved drugs, and varies dramatically among disease areas. We estimate that selecting genetically supported targets could double the success rate in clinical development. Therefore, using the growing wealth of human genetic data to select the best targets and indications should have a measurable impact on the successful development of
    Nature Genetics 06/2015; advance online publication. DOI:10.1038/ng.3314 · 29.35 Impact Factor
  • Jiang Li · Pak Chung Sham · Youqiang Song · Miaoxin Li ·
    [Show abstract] [Hide abstract]
    ABSTRACT: Set-based association tests, combining a set of single-nucleotide polymorphisms into a unified test, have become important approaches to identify weak-effect or low-frequency risk loci of complex diseases. However, there is no comprehensive and user-friendly tool to estimate power of set-based tests for study design. We developed a simulation tool to estimate statistical power of multiple representative set-based tests (SPS). SPS has a graphic interface to facilitate parameter settings and result visualization. Advanced functions include loading real genotypes to define genetic architecture, set-based meta-analysis for risk loci with or without heterogeneity, and parallel simulations. In proof-of-principle examples, SPS took no more than 3 sec on average to estimate the power in a conventional setting. The SPS has been integrated into a user-friendly software tool (KGG) as an independent functional module and it is freely available at © 2015 WILEY PERIODICALS, INC.
    Genetic Epidemiology 05/2015; 39(5). DOI:10.1002/gepi.21898 · 2.60 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent founder mutations may play important roles in complex diseases and Mendelian disorders. Detecting shared haplotypes that are identical by descent (IBD) could facilitate discovery of these mutations. Several programs address this, but are usually limited to detecting pair-wise shared haplotypes and not providing a comparison of cases and controls. We present a novel algorithm and software package, HaploShare, which detects extended haplotypes that are shared by multiple individuals, and allows comparisons between cases and controls. Testing on simulated and real cases demonstrated significant improvements in detection power and reduction of false positive rate by HaploShare relative to other programs. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0662-9) contains supplementary material, which is available to authorized users.
    Genome biology 05/2015; 16(1):92. DOI:10.1186/s13059-015-0662-9 · 10.81 Impact Factor

  • Maturitas 05/2015; 81(1):172. DOI:10.1016/j.maturitas.2015.02.211 · 2.94 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Epilepsy is a common neurological disease characterized by recurrent unprovoked seizures. Evidence suggested that abnormal activity of brain-derived neurotrophic factor (BDNF) contributes to the pathogenesis of epilepsy. Some previous studies identified association between genetic variants of BDNF and risk of epilepsy. In this study, this association has been examined in the Hong Kong and Malaysian epilepsy cohorts. Genomic DNA of 6047 subjects (1640 patients with epilepsy and 4407 healthy individuals) was genotyped for rs6265, rs11030104, rs7103411 and rs7127507 polymorphisms by using Sequenom MassArray and Illumina HumanHap 610-Quad or 550-Duo BeadChip arrays techniques. Results showed significant association between rs6265 T, rs7103411 C, and rs7127507 T and cryptgenic epilepsy risk (p = 0.00003, p = 0.0002, and p = 0.002, respectively) or between rs6265 and rs7103411 and symptomatic epilepsy risk in Malaysian Indians (TT vs. CC, p = 0.004 and T vs. C, p = 0.0002, respectively) as well as between rs6265 T and risk of cryptogenic epilepsy in Malaysian Chinese (p = 0.005). The Trs6265-Crs7103411-Trs7127507 was significantly associated with cryptogenic epilepsy in Malaysian Indians (p = 0.00005). In concusion, our results suggest that BDNF polymorphisms might contribute to the risk of epilepsy in Malaysian Indians and Chinese.
    Molecular Neurobiology 04/2015; DOI:10.1007/s12035-015-9150-1 · 5.14 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Genetic interaction has been considered as a hallmark of the genetic architecture of systemic lupus erythematosus (SLE). Based on two independent genome-wide association studies (GWAS) on Chinese populations, we performed a genome-wide search for genetic interactions contributing to SLE susceptibility. The study involved a total of 1 659 cases and 3 398 controls in the discovery stage and 2 612 cases and 3 441 controls in three cohorts for replication. Logistic regression and multifactor dimensionality reduction were used to search for genetic interaction. Interaction of CD80 (rs2222631) and ALOX5AP (rs12876893) was found to be significantly associated with SLE (OR_int=1.16, P_int_all=7.7E-04 at false discovery rate<0.05). Single nuclear polymorphism rs2222631 was found associated with SLE with genome-wide significance (P_all=4.5E-08, OR=0.86) and is independent of rs6804441 in CD80, whose association was reported previously. Significant correlation was observed between expression of these two genes in healthy controls and SLE cases, together with differential expression of these genes between cases and controls, observed from individuals from the Hong Kong cohort. Genetic interactions between BLK (rs13277113) and DDX6 (rs4639966), and between TNFSF4 (rs844648) and PXK (rs6445975) were also observed in both GWAS data sets. Our study represents the first genome-wide evaluation of epistasis interactions on SLE and the findings suggest interactions and independent variants may help partially explain missing heritability for complex diseases. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to
    Annals of the rheumatic diseases 04/2015; DOI:10.1136/annrheumdis-2014-206367 · 10.38 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Systemic lupus erythematosus (SLE) is a heterogeneous disease with a diverse spectrum of clinical symptoms from skin rash to end-organ damage. 22q11.21 has been identified as a susceptibility region for several autoimmune diseases, including SLE. However, the detailed information for SLE association and the underlying functional mechanism(s) are still lacking. Through meta-analysis of two genome-wide association studies (GWAS) on Chinese Han populations with a total of 1659 cases and 3398 controls matched geographically, we closely examined this region, especially on the reported single nucleotide polymorphisms (SNPs) associated with different autoimmune diseases and their relationships. We further replicated the most significant association SNP with SLE using 2612 cases and 2323 controls of Asian ancestry. All reported SNPs in this region with different autoimmune diseases were examined in the two GWAS data and meta- analysis result, and supportive evidence of association with SLE was found (meta-analysis P_meta ≤ 7.27E-05), which might require further investigation. SNP rs2298428 was identified as the most significant SNP associated with SLE in this region (P_meta = 2.70E-09). It showed independent effect through both stepwise and conditional logistic regression, and there is no evidence of other independent association signals for SLE in this region. The association of rs2298428 was further replicated in three cohorts from Hong Kong, Anhui and Thailand with a total of 2612 cases and 2323 controls (joint analysis of GWAS and replication result P_all = 1.31E-11, OR = 1.23). SNP rs2298428 was shown to be an eQTL for UBE2L3 gene in different cell types, with the risk allele (T) being correlated with higher expression of UBE2L3. This is consistent with earlier reports on higher expression of UBE2L3 in SLE cases. Association to distinct autoimmune diseases highlights the significance of this region in autoreactive responses and potentially shared functional mechanisms by these diseases.
    Arthritis research & therapy 03/2015; 17(1):67. DOI:10.1186/s13075-015-0577-6 · 3.75 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Human leukocyte antigen (HLA) typing from next generation sequencing (NGS) data has the potential for widespread applications. Here we introduce a novel tool (HLAreporter) for HLA typing from NGS data based on read-mapping using a comprehensive reference panel containing all known HLA alleles, followed by de novo assembly of the gene-specific short reads. Accurate HLA typing at high-digit resolution was achieved when it was tested on publicly available NGS data, outperforming other newly developed tools such as HLAminer and PHLAT. HLAreporter can be downloaded from Electronic supplementary material The online version of this article (doi:10.1186/s13073-015-0145-3) contains supplementary material, which is available to authorized users.
    Genome Medicine 03/2015; 7(1):25. DOI:10.1186/s13073-015-0145-3 · 5.34 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The fatality of avian influenza A(H7N9) infection in humans was over 30%. To identify human genetic susceptibility to A(H7N9) infection, we performed a genome-wide association study (GWAS) involving 102 A(H7N9) patients and 106 heavily-exposed healthy poultry workers, a sample size critically restricted by the small number of human A(H7N9) cases. To tackle the stringent significance cutoff of GWAS, we utilized an artificial imputation program SnipSnip to improve the association signals. In single-SNP analysis, one of the top SNPs was rs13057866 of LGALS1. The artificial imputation (AI) identified three non-genotyped causal variants, which can be represented by three anchor/partner SNP pairs rs13057866/rs9622682 (AI P = 1.81 × 10−7), rs4820294/rs2899292 (2.13 × 10−7) and rs62236673/rs2899292 (4.25 × 10−7) respectively. Haplotype analysis of rs4820294 and rs2899292 could simulate the signal of a causal variant. The rs4820294/rs2899292 haplotype GG, in association with protection from A(H7N9) infection (OR = 0.26, P = 5.92 × 10−7) correlated to significantly higher levels of LGALS1 mRNA (P = 0.050) and protein expression (P = 0.025) in lymphoblast cell lines. Additionally, rs4820294 was mapped as an eQTL in human primary monocytes and lung tissues. In conclusion, functional variants of LGALS1 causing the expression variations are contributable to the differential susceptibility to influenza A(H7N9).
    Scientific Reports 02/2015; 5:8517. DOI:10.1038/srep08517 · 5.58 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although expression quantitative trait locus, eQTL, serves as an explicit indicator of gene-gene associations, challenges remain to disentangle the mechanisms by which genetic variations alter gene expression. Here we combined eQTL and molecular analyses to identify an association between two seemingly non-associated genes in brain expression data from BXD inbred mice, namely Ptpn21 and Nrg3. Using biotinylated receptor tracking and immunoprecipitation analyses, we determined that PTPN21 de-phosphorylates the upstream receptor tyrosine kinase ErbB4 leading to the up regulation of its downstream signaling. Conversely, kinase-dead ErbB4 (K751R) or phosphatase-dead PTPN21 (C1108S) mutants impede PTPN21-dependent signaling. Furthermore, PTPN21 also induced Elk-1 activation in embryonic cortical neurons and a novel Elk-1 binding motif was identified in a region located 1919bp upstream of the NRG3 initiation codon. This enables PTPN21 to promote NRG3 expression through Elk-1, which provides a biochemical mechanism for the PTPN21-NRG3 association identified by eQTL. Biologically, PTPN21 positively influences cortical neuronal survival and, similar to Elk-1, it also enhances neuritic length. Our combined approaches show for the first time, a link between NRG3 and PTPN21 within a signalling cascade. This may explain why these two seemingly unrelated genes have previously been identified as risk genes for schizophrenia. Copyright © 2015. Published by Elsevier Ltd.
    The International Journal of Biochemistry & Cell Biology 02/2015; 61. DOI:10.1016/j.biocel.2015.02.003 · 4.05 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: With the rapid advances in high-throughput sequencing technologies, exome sequencing and targeted region sequencing have become routine approaches for identifying mutations of inherited disorders in both genetics research and molecular diagnosis. There is an imminent need for comprehensive and easy-to-use downstream analysis tool to isolate causal mutations in exome sequencing studies. We have developed a user-friendly online framework, wKGGSeq, to provide systematic annotation, filtration, prioritization and visualization functions for characterizing causal mutation(s) in exome sequencing studies of inherited disorders. wKGGSeq provides: 1) a novel strategy-based procedure for downstream analysis of a large amount of exome sequencing data and 2) a disease-targeted analysis procedure to facilitate clinical diagnosis of well-studied genetic diseases. In addition, it is also equipped with abundant online annotation functions for sequence variants. We demonstrate that wKGGSeq either outperforms or is comparable to two popular tools in several real exome sequencing samples. This tool will greatly facilitate the downstream analysis of exome sequencing data and can play a useful role for researchers and clinicians in identifying causal mutations of inherited disorders. The wKGGSeq is freely available at or, and will be updated frequently. This article is protected by copyright. All rights reserved. This article is protected by copyright. All rights reserved.
    Human Mutation 02/2015; 36(5). DOI:10.1002/humu.22766 · 5.14 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: High throughput assays tend to be expensive per subject. Often studies are limited not so much by the number of subjects available as by assay costs, making assay choice a critical issue. We have developed a framework for assay choice that maximises the number of true disease causing mechanisms `seen¿, given limited resources. Although straightforward, some of the ramifications of our methodology run counter to received wisdom on study design. We illustrate our methodology with examples, and have built a website allowing calculation of quantities of interest to those designing rare disease studies.
    Orphanet Journal of Rare Diseases 02/2015; 10(1):10. DOI:10.1186/s13023-015-0226-9 · 3.36 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Thoracic tumor, especially lung cancer, ranks as the top cancer mortality in most parts of the world. Lung adenocarcinoma is the predominant subtype and there is increasing knowledge on therapeutic molecular targets, namely EGFR, ALK, KRAS, and ROS1, among lung cancers. Lung cancer cell lines established with known clinical characteristics and molecular profiling of oncogenic targets like ALK or KRAS could be useful tools for understanding the biology of known molecular targets as well as for drug testing and screening. Five new cancer cell lines were established from pleural fluid or biopsy tissues obtained from Chinese patients with primary lung adenocarcinomas or malignant pleural mesothelioma. They were characterized by immunohistochemistry, growth kinetics, tests for tumorigenicity, EGFR and KRAS gene mutations, ALK gene rearrangement and OncoSeq mutation profiling. These newly established lung adenocarcinoma and mesothelioma cell lines were maintained for over 100 passages and demonstrated morphological and immunohistochemical features as well as growth kinetics of tumor cell lines. One of these new cell lines bears EML4-ALK rearrangement variant 2, two lung cancer cell lines bear different KRAS mutations at codon 12, and known single nucleotide polymorphism variants were identified in these cell lines. Four new lung adenocarcinoma and one mesothelioma cell lines were established from patients with different clinical characteristics and oncogenic mutation profiles. These characterized cell lines and their mutation profiles will provide resources for exploration of lung cancer and mesothelioma biology with regard to the presence of known oncogenic mutations.
    OncoTargets and Therapy 01/2015; 8:195-209. DOI:10.2147/OTT.S71242 · 2.31 Impact Factor
  • Source

Publication Stats

37k Citations
3,064.97 Total Impact Points


  • 2009-2015
    • Lands Department of The Government of the Hong Kong Special Administrative Region
      Hong Kong, Hong Kong
  • 2004-2015
    • The University of Hong Kong
      • • Centre for Genomic Sciences
      • • Department of Psychiatry
      • • State Key Laboratory of Brain and Cognitive Sciences
      • • Department of Medicine
      Hong Kong, Hong Kong
  • 2004-2014
    • Queen Mary Hospital
      Hong Kong, Hong Kong
  • 2011
    • The Hong Kong Polytechnic University
      • Department of Health Technology and Informatics
      Hong Kong, Hong Kong
  • 1992-2008
    • King's College London
      • • MRC Social, Genetic and Developmental Psychiatry Centre
      • • Department of Psychological Medicine
      • • Institute of Psychiatry
      Londinium, England, United Kingdom
  • 2006
    • National University of Ireland, Galway
      • Department of Psychiatry
      Gaillimh, Connaught, Ireland
    • Toronto Western Hospital
      Toronto, Ontario, Canada
  • 2004-2006
    • University of Cambridge
      • Department of Psychiatry
      Cambridge, England, United Kingdom
  • 1999-2006
    • London Research Institute
      Londinium, England, United Kingdom
    • University of London
      Londinium, England, United Kingdom
    • University of Missouri
      Columbia, Missouri, United States
    • Sapienza University of Rome
      Roma, Latium, Italy
  • 2005
    • Queen Mary, University of London
      • Centre for Psychiatry
      Londinium, England, United Kingdom
    • Shanghai Jiao Tong University
      • Bio-X Institute
      Shanghai, Shanghai Shi, China
  • 1996-2005
    • Institute of Genetics and Molecular Medicine
      Edinburgh, Scotland, United Kingdom
    • Clinique médicale et pédagogique Dupré
      Île-de-France, France
  • 1998-2004
    • The Kings College
      Колумбия, Tennessee, United States
  • 2003
    • Chang Gung Memorial Hospital
      • Department of Psychiatry
      T’ai-pei, Taipei, Taiwan
    • Fundación Argibide
      Iruña, Navarre, Spain
    • University College London
      • Department of Epidemiology and Public Health
      London, ENG, United Kingdom
  • 1997-2003
    • Maastricht University
      • Department of Epidemiology
      Maastricht, Provincie Limburg, Netherlands
    • Teikyo University
      Edo, Tōkyō, Japan
  • 2002
    • University of South Wales
      Понтиприте, Wales, United Kingdom
  • 2000-2002
    • Instituto Nacional de Psiquiatría
      Ciudad de México, Mexico City, Mexico
    • Sichuan University
      Hua-yang, Sichuan, China
  • 2001
    • UK Energy Research Centre
      Londinium, England, United Kingdom
  • 1995-1999
    • New South Wales Institute of Psychiatry
      Sydney, New South Wales, Australia
    • University of Essex
      Colchester, England, United Kingdom
    • Teikyo University Hospital
      Edo, Tōkyō, Japan
    • St. James's Hospital
      Dublin, Leinster, Ireland
  • 1993-1995
    • Virginia Commonwealth University
      • Department of Psychiatry
      Ричмонд, Virginia, United States
  • 1991
    • University of Nottingham
      • Division of Psychiatry and Applied Psychology
      Nottigham, England, United Kingdom