Original Article

Genes and Immunity (2010) 11, 232–238; doi:10.1038/gene.2010.1; published online 18 March 2010

New genetic associations detected in a host response study to hepatitis B vaccine

S Davila1,5, F E M Froeling2,5,6, A Tan3, C Bonnard4, G J Boland2, H Snippe2, M L Hibberd1 and M Seielstad4

  1. 1Infectious Diseases Group, Genome Institute of Singapore, Singapore
  2. 2Department of Medical Microbiology and Virology, University Medical Center Utrecht, Utrecht, The Netherlands
  3. 3Research Computing Group, Genome Institute of Singapore, Singapore
  4. 4Human Genetics Group, Genome Institute of Singapore, Singapore

Correspondence: Dr M Seielstad, Population Genetics, Genome Institute of Singapore, 60 Biopolis St., Singapore 138672, Singapore. E-mail: seielstadm@gis.a-star.edu.sg; Dr S Davila, Infectious Diseases, Genome Institute of Singapore, 60 Biopolis St., Singapore 138672, Singapore. E-mail: sonia@gis.a-star.edu.sg

5These authors contributed equally to this work.

6Current address: Centre for Tumour Biology, Institute of Cancer, Barts and the London School of Medicine and Dentistry, London EC1M 6BQ, UK.

Received 16 June 2009; Revised 1 December 2009; Accepted 1 December 2009; Published online 18 March 2010.



The immune response to hepatitis B vaccination differs greatly among individuals, with 5–10% of healthy people failing to produce protective levels of antibodies. Several factors have been implicated in determining this response, chiefly individual genetic variation and age. Aiming to identify genes involved in the response to hepatitis B vaccination, a two-stage investigation of 6091 single-nucleotide polymorphisms (SNPs) in 914 immune genes was performed in an Indonesian cohort of 981 individuals showing normal levels of anti-HBs versus 665 individuals displaying undetectable levels of anti-HBs 18 months after initial dose of the vaccine. Of 275 SNPs identified in the first stage (476 normal/372 nonresponders) with P<0.05, significant associations were replicated for 25 polymorphisms in 15 genes (503 normal/295 nonresponders). We validated previous findings (HLA-DRA, rs5000563, P-value combined=5.57 × 10−10; OR (95%CI)=0.61 (0.52–0.71)). In addition, we detected a new association outside of the human leukocyte antigen loci region that passed correction for multiple testing. This SNP is in the 3′ downstream region of FOXP1, a transcription factor involved in B-cell development (P-value combined=9.2 × 10−6; OR (95%CI)=1.38 (1.2–1.6)).These findings might help to understand the biological reasons behind vaccine failure and other aspects of variation in the immune responses of healthy individuals.


hepatitis B; vaccines; human genetics; major histocompatibility complex



More than 2 billion people have been infected with the hepatitis B virus (HBV), of which more than 350 million are chronic carriers. Each year more than half a million people die as a result of acute or chronic HBV infection.1 Up to 90% of infants infected in their first year of life and about 10% of exposed, immunocompetent adults develop chronic infection, placing them at high risk of developing progressive liver disease, cirrhosis and hepatocellular carcinoma because of continual viral replication in hepatocytes.2 Vaccination has been highly successful in preventing new HBV infections and has been implemented into the national immunization programs of more than 150 countries. However, 5–10% of healthy adults fail to produce protective levels of antibodies against the hepatitis B surface antigen (anti-HBs) contained in the vaccine.3 The reasons for these individual failures are multiple, including physical factors such as age, gender, obesity, immunosuppression and smoking4, 5 as well as variation in genes of the immune system.6, 7, 8, 9

Several polymorphisms of the human leukocyte antigen (HLA) loci have been linked to variations in the immune response both to vaccination as well as natural hepatitis B infections. In particular, HLA-DR3 and DR7, which are in linkage disequilibrium with HLA-DQ2, have been shown to be associated with nonresponsiveness,10, 11 whereas DRB1* alleles 1, 11 and 15 have been shown to be associated with a good response to HBV vaccination.7 Although an association with these HLA class II genotypes has become increasingly evident, case–control studies that analyzed potential non-major histocompatibility complex (MHC) candidate genes have so far yielded conflicting results or have not been confirmed by other investigators.12, 13, 14, 15, 16

To date, most studies have examined the function of just one or two genes in the immune response to HBV. A more extensive approach has recently been taken by Hennig et al.17 who studied the association of 715 single-nucleotide polymorphisms (SNPs), across 133 candidate genes, measuring peak vaccine induced anti-HBs level and anti-HBc status in 662 infant vaccinees from the Gambia. The three most strongly associated genes identified in their first screen were MAPK8 (MIM: 601158), IFNG (MIM: 147570) and IL10RA (MIM: 146933), suggesting that genetically determined differences in interferon-γ (IFN-γ) activity, intereukin-10 (IL10) receptor variation and factors affecting T-cell differentiation influence the response to the hepatitis B vaccine. However, the only polymorphism that could be replicated in a second screen was a nonsynonymous coding change in ITGAL (MIM: 153370), a gene with a central function in immune cell interaction, which was shown to be associated with increased peak anti-HBs levels.

Recent developments in high-throughput genotyping technology have enabled the genotyping of hundreds of thousands of SNPs throughout the genome.18, 19 However, association studies that test a large number of SNPs can lead to false-positive associations if multiple testing is not adequately accounted for.20 At the same time, correcting for tens or hundreds of thousands of tests requires large and sometimes unrealistic sample sizes to maintain adequate statistical power while minimizing the accompanying risk of false-negative associations. To balance these costs and benefits, and because immune-related genes have such a strong a priori probability of involvement in a variable immune response, we have adopted a focused approach to examine variation in 914 genes with a direct function in immunity and host defense.

Between 1996 and 2002, a long-term catch-up vaccination study was performed in Indonesia, investigating the protective efficacy of a two-dose vaccination schedule.4 Here, we examine genetic variation in 914 immune candidate genes in participants of this study to implicate genetic variation that may have a function in determining postvaccination antibody titer, and perhaps, by extension, variation affecting the response to the HBV itself.



Demographic details of patients and controls genotyped in both stages are summarized in Table 1. As mentioned before, the two groups were matched for gender. All individuals genotyped resided on the Indonesian island of Batam or its immediate neighbors. Using principal components analysis, we did not detect evidence for significant population stratification in our study group. When compared to HapMap populations, the Indonesian group clustered fully within the two Asian populations, Japanese from Tokyo and Chinese from Beijing (Supplementary Figure S1).

Out of 935 samples genotyped in the first stage, 13 were excluded due to detection of a gender discrepancy between recorded data and genetic information. Samples with call rates below 90%, N=23, were also removed. Fifty-one samples were removed after detection of a close family relationship. Thus, 848 samples (372 cases and 476 controls) passed all quality control filters and were included in the study.

A total of 6091 SNPs were genotyped using an Illumina GoldenGate assay (Illumina Inc., San Diego, CA, USA), of which 391 had call rates below 90%, 12 had minor allele frequencies below 0.005 and 56 were out of Hardy–Weinberg equilibrium in controls after applying a Bonferroni correction. Hence, 5632 SNPs were analyzed in the first stage (Supplementary Table S1). A trend test identified 298 SNPs with P-values below 0.05. However, after performing a logistic regression adjusting for age, a known nongenetic component of the host response to hepatitis B vaccine,5 the number of significant SNPs dropped to 275 (Supplementary Table S2).

Genotyping of the second-stage validation cohort consisting of 842 samples for the 275 SNPs carried forward from stage 1 was conducted with a combination of two different platforms (described in Subjects and methods section). Six samples showed a gender discrepancy and 35 were excluded due to detection of family relationships. Three samples had call rates below 90%. Hence, 798 samples (295 cases and 503 controls) were included in subsequent analyses of the 235 SNPs attempted on the GoldenGate platform. Ten SNPs genotyped by Illumina GoldenGate, did not pass quality control, leaving 225 SNPs for association analysis. For the SNPs typed on the Sequenom iPLEX platform (Sequenom, San Diego, CA, USA), 76 samples failed genotyping, leaving 277 cases and 438 controls to be analyzed for the 40 SNPs typed by this method. Of the 40 SNPs attempted by iPLEX, two failed the assay design process, and three failed during genotyping.

Of the 260 successfully genotyped SNPs in stage 2, 25 attained P-values below 0.05 in the replication study after correcting for age (Table 2). Of these, 13 were intragenic, including 4 nonsynonymous polymorphisms and 2 in the MHC region. None of the associated SNPs showed a significant P-value for the Breslow–Day test, implying a lack of heterogeneity in effect size among the two stages (Supplementary Table S3). After applying a permutation test for correction for multiple testing seven SNPs maintained their statistical significance (Table 2). Consistent with previous reports,21 we detected a strong genetic association between host response to hepatitis B vaccine and the MHC locus on chromosome 6 (rs5000563, most significant combined P-value=5.6 × 10–10; OR (95%CI)=0.60 (0.52–0.71)). Up to 11 polymorphisms within this region, spanning over 1.3Mb, showed significant association in the second stage. Two of the SNPs are predicted to alter the protein sequence of BTNL2, rs2076523 (Lys196Glu), and HLA-DRA, rs7192 (Leu242Val) (Table 2).

Beyond the MHC, one of the SNPs that decreases host response to HBV vaccine and passed multiple testing correction was located downstream of the transcription factor FOXP1 (MIM: 605515), rs6789153 (P-value combined=9.2 × 10–6; OR (95%CI)=1.38 (1.20–1.60)) (Table 2).

A polymorphism, rs1654668, situated in the 5′ untranslated region (UTR) of leukocyte immunoglobulin-like receptor, subfamily B, member 4 (LILRB4, MIM: 604821, also known as ILT3), also achieved a small P-value associated with bad response (P-value combined=8 × 10–5; OR (95%CI)=1.34 (1.16–1.56)) (Table 2).

Several genes had more than one associated polymorphism. Among them (1) two intronic SNPs within tumor necrosis factor ligand superfamily 15 (TNFSF15, MIM: 604052) (P-value combined=4.8 × 10–3–5.8 × 10–3; OR (95%CI)=1.22 (1.06–1.41) to 1.22 (1.06–1.41)), (2) complement component 5 (C5, MIM: 120900), also had two associated SNPs, rs1978270 and rs7029078, located 15kb apart in intronic regions (P-value combined=1.1 × 10–2 to 8.4 × 10–3; OR (95%CI)=1.3 (1.06–1.60) to 1.3 (1.07–1.60), respectively) and (3) the chemokine (C-C motif) ligand 15 (CCL15, MIM: 601393) with one intronic, rs854692, and one nonsynonymous SNP, rs854625 (Ile24Thr) (P-value combined=9.6 × 10–4 to 1.2 × 10–3; OR (95%CI)=1.28 (1.11–1.50) to 1.29 (1.11–1.51), respectively) (Table 2).

Two genes from the same family were implicated, transforming growth factor, β2 (TGFB2, MIM: 190220), and transforming growth factor, β3 (TGFB3, MIM: 190230). These genes are located on different chromosomes, and the detected polymorphisms were replicated in the second stage. In both cases, the SNPs were associated with a good response (OR (95%CI)=0.72 (0.62–0.85) to 0.79 (0.68–0.92), respectively) (Table 2). A nonsynonymous polymorphism, rs2232548 (Leu67Phe), in killer cell lectin-like receptor subfamily F, member 1 (KLRF1, MIM: 605029), was also found to be associated with vaccine responsiveness (P-value combined=3.5 × 10–4; OR (95%CI)=0.59 (0.45–0.79)).



In this study we aimed to identify genes involved in the host response to hepatitis B vaccine in an Indonesian cohort. Several reports have associated genetic variability in the HLA locus with host response to hepatitis B vaccine in different populations.21, 22, 23 Consistent with those studies, we detected a strong genetic association with the HLA region in the current Indonesian population sample. Out of 64 genotyped SNPs within the HLA region, 16 showed nominally significant P-values in the first stage (Supplementary Table S1). Of these 16, 11 were replicated in the second stage confirming the initial association. Six survived correction for multiple testing (Table 2). Although most of the significant SNPs are in or near the HLA-DR region, which has been the focus of previous studies, the HLA region shows particularly complex and extensive patterns of linkage disequilibrium encompassing numerous genes involved in immunity.24 A fine mapping strategy including human histocompatibility leukocyte antigen (HLA) typing may help to identify the functionally relevant alleles, though this task was beyond the scope of our project. Nevertheless, we believe that replication of this previously reported and highly plausible association indicates the validity of our experimental approach and population sample. A recent genome-wide association study reported association of the HLA-DP locus with chronic hepatitis B infection in Asians,25 highlighting once again the importance of the highly polymorphic HLA locus in the host immune response to HBV. In our study we did not genotype any of the significant SNPs detected by Kamatani et al.,25 however, our population showed a significant association with HLA-DQB and DRA loci located 5′ upstream of HLA-DP. All of them belong to the MHC class II molecules whose main function is to present antigens to CD4+ T cells, leading to suppression or stimulation of antibody production against the presented antigen. Hence, one supposes that HLA alleles that fail to recognize or that bind only weakly to the s-antigen fragment found in the vaccine are associated with a poor vaccine response. Identifying such alleles could help guide the design of a next-generation vaccine containing peptides that such HLA alleles would be capable of binding and presenting to T cells.

It has been shown that genes outside the MHC region also have an important function in host response to hepatitis B vaccine, accounting for more than half of the genetic effects.7, 9 One non-MHC gene emerges with highly compelling evidence of support in our study, FOXP1. FOXP1 is a transcription factor that is instrumental in regulating development of the antibody producing B cell.26 It is encoded by a large gene, located on chromosome 3, consisting of 21 exons. The associated SNP, rs6789153, was located more than 15kb away from the closest genotyped polymorphism within the gene itself, and was not in linkage disequilibrium with any of the typed SNPs within the gene. Additional study of the 3′ region of the gene should help to establish a more definitive function of FOXP1 in the host response to the vaccine.

Two other genes outside of the MHC region showed consistent evidence of association, although they did not pass correction for multiple testing, LILRB4 and TGFB2. An SNP within the LILRB4 gene, rs1654668, was found to be associated with poor responsiveness (OR=1.34 (1.16–1.56)). LILRB4 encodes for a receptor protein expressed by dendritic cells, monocytes, macrophages and endothelial cells that on binding of MHC class I molecules triggers T-cell inhibition.27, 28 This effect of suppressing inflammation has been shown in vivo in a mouse model.29 In humans an inverse correlation between high expression of LILRB4 in patients that underwent organ transplant and development of chronic rejection has been reported,30 which supports the anti-inflammatory effect of LILRB4. TGFB is a cytokine that has been shown to be involved in a number of processes in the liver, such as fibrogenesis and pro-apoptosis.31 Its function on neoplasia during hepatocellular carcinoma remains contentious, but a recent paper has reported an increase in DNA copy number of TGFB2 in hepatocellular carcinoma infected with hepatitis C virus.32

We have found that three genes outside of the MHC region, TNFSF15, C5 and CCL15, harbor more than one significantly associated polymorphism (Table 2). Linkage disequilibrium among SNPs in two of these genes is high in our population, TNFSF15, r2=0.95 (Supplementary Figure S2) and C5, r2=0.88 (Supplementary Figure S3). These results agree with the HapMap data from Asian populations (r2=0.97 in both cases), which diminishes the possibility of spurious association due to genotyping errors and implies that both association signals arise from the same, single, functional allele. Three SNPs were found within complement component C5 in the first stage of the study (Supplementary Table S2). Two of them were replicated in the second stage. C5 encodes for a protein known to be fundamental in inflammatory and cell killing processes. A polymorphism within the TRAF1/C5 region has been associated with susceptibility to rheumatoid arthritis.33 Interestingly, two of the genes identified in this study, C5 and LILRB4, suggest that polymorphisms can disrupt a fine balance between the survival benefits of an acute inflammatory response to infection, and the destructive effects of chronic inflammation. As for CCL15, we did not observe a high degree of LD among the associated SNPs. However, a group of 10 SNPs on chromosome 17, where a cluster of cytokines is located, reached significant P-values in the first stage (Supplementary Table S2). Cytokines are secreted proteins involved in immune and inflammatory responses. Two of the SNPs cause nonsynonymous changes in the protein sequence, rs854625 in CCL15 (Ile24Thr), and rs1003645 in CCL23 (MIM: 602494) (Met123Val), though the effects of these polymorphisms on protein function have not yet been established, but should warrant further investigation. Five polymorphisms were replicated in the second stage, although the statistical significance of three of them was reduced after correcting for nongenetic factors such age. These results indicate a likely function for cytokines in host response to hepatitis B vaccine.

Detection of polymorphisms located outside of the HLA region that influence response to HBV vaccination has been reported recently by Hennig et al.17 They identified six non-HLA genes associated with anti-HBs levels and anti-HBc status. Two of them, MAPK8 and ITGAL, were not included in our study. Two SNPs within CD58 (rs1414275, rs1016140) and one within CD163 (rs4883263) were genotyped in both studies; however, they did not reach any measure of significance in our population (Supplementary Table 1). Only one polymorphism within IFNG showed a significant value (rs2069718, P=0.012) in our first stage, although without any replication in our second stage (P=0.4). Although this SNP shows an r2=1.0 with the significant SNP (rs2069727) reported by Hennig et al. in the HapMap Asian populations, in the HapMap Yoruba, r2 drops to only 0.29. Thus the potential function of variation in IFNG and hepatitis B vaccine response may deserve more attention in Asian populations. Considering the poor degree of replication between these two studies overall, there are several potential explanations: (1) their study was carried out in an African population whereas we reported results from an Asian group, (2) most of their subjects were below 20 years of age, whereas in our study the age ranged from 5 to 76 years and (3) both reports looked at different outcomes: peak anti-HBs levels and anti-HBc status by Hennig et al. and only antibody titer (anti-HBs) 6 months after the second vaccination in this work.

Here we report seven polymorphisms, six within and one outside of the HLA region, that achieved statistical significance after applying correction for multiple testing in a genetic association study of hepatitis B vaccine response in an Indonesian population. Identification of non-HLA genes likely to account for up to 50% of heritability to hepatitis B vaccine response7 might help us to design more effective vaccines that will then decrease the percentage of nonresponders or low responders in the population.


Subjects and methods

HBV vaccine study group

Our study sample came from a previous trial aimed at assessing the efficacy of a two-dose hepatitis B vaccination scheme, conducted on the Indonesian island of Batam. Study design, eligibility criteria, demographic characteristics and methods for virologic and immunologic measurements have been reported elsewhere.4 This study initially screened 19547 individuals over 5 years of age. Subjects who were anti-HBc, anti-HBs or HBsAg positive at entry (N=4802) were excluded. A two-dose vaccination schedule (Engerix-B, 20μgl−1 recombinant HBsAg; GlaxoSmithKline Biologicals, Rixensart, Belgium) was completed in 13315 subjects. To recipients of the study, the two doses were given in both the arms with either a 6- or 12-month interval. After 18 months, antibody titer was measured in 5025 subjects (Elecsys system; Boehringer/Roche, Mannheim, Germany). No response (that is, undetectable anti-HBs) was seen in 811 subjects, a low response (that is, anti-HBs 10–100IUl−1) was detected in 773 subjects, 1495 subjects had a normal response (that is, anti-HBs levels in the range of 100–1000IUl−1) and 1946 subjects showed a high response (that is, anti-HBs >1000IUl−1) to the vaccine.3, 34

The study was approved by the medical ethical review board of the Utrecht Medical Center. Consent forms were signed on recruitment by all participants. Ethical approval was obtained again in May 2006 for testing genetic factors influencing the vaccine response.

Sample selection

Sample characteristics were evaluated with the statistical software SPSS for Windows version 12.0.1 (release 2003; SPSS Inc., Chicago, IL, USA). For the first stage, 405 cases (that is, nonresponders, undetectable anti-HBs) were selected randomly and 530 controls (that is, normal responders, anti-HBs levels in the range of 100–1000IUl−1) were chosen to match for gender against cases. An additional 304 cases and 538 controls were included in the second stage. The two groups were matched for gender, and a significant difference in age persisted in both stages after excluding samples that did not pass quality controls (Table 1). Differences in vaccination schedule did not contribute to vaccine response.4

Sample duplications and family relationships were identified using PLINK.35 Pairs with an identity by descent proportion greater than or equal to 0.4 were considered to be first-degree relatives. In such instances, one member of each relative pair was dropped from the analysis (25 cases and 61 controls), retaining cases over controls or the sample with higher call rate if both belonged to the same group.

Principal components analysis36 was used to check for signs of population stratification. No sample fell beyond three standard deviations from the mean in a principal components analysis of Asian HapMap populations, Chinese from Beijing and Japanese from Tokyo (Supplementary Figure S1). Thus, no samples were removed as population outliers and no further adjustments for population stratification were made.

DNA amplification

DNA was isolated from frozen serum samples of 1068 normal responders and 709 nonresponders using standard protocols. Whole-genome amplification of DNA was performed by GenomePlex Technology (Rubicon Genomics Inc., Ann Arbor, MI, USA).37 Of the 1777 total subjects selected, 1774 (99.9%) provided suitable DNA specimens for genetic analyses.

Candidate gene and SNP selection

We adopted a two-stage study design38, 39 in which a set of SNPs attaining a relatively liberal P-value in an initial case–control series (stage 1) was carried forward for genotyping in a second similar series of cases and controls (stage 2). Replication of significant P-values in the second stage and combined tests of significance across both stages were used to distinguish the few true-positive associations identified in stage 1 from the many false-positive results that occur by chance or as a result of population stratification or other biases in study design.

For stage 1, genotypes were attempted for a total of 6091 SNPs from 914 candidate genes. These genes were manually selected from the approximately 1700 genes with a direct function in the immune response after filtering for biological process and molecular function based on the PANTHER Classification System.40 SNPs within each gene were chosen based on their likely functional importance, a minor allele frequency in Asian populations exceeding 5%, and for evenness of coverage. SNPs were identified using an in-house database, GISSNP that integrates data from public databases (Ensembl, Celera, NCBI) as well as proprietary data. A full list of the 5632 successfully genotyped SNPs is provided in Supplementary Table S1.

SNP genotyping

Genotyping of 6091 SNPs was attempted according to manufacturer's protocols using GoldenGate assays on an Illumina BeadStation 500G (Illumina Inc.).41 To minimize bias in genotyping errors, we included a mix of samples from responders and nonresponders in each plate. SNPs and samples with call rates below 90% were excluded from further analysis.

SNPs with P-values <0.05 in the Cochran–Armitage test for trend were selected for follow-up in the second stage. Genotyping of 252 SNPs was performed with an Illumina GoldenGate assay. The Sequenom iPLEX protocol.42 was used for the genotyping of 46 SNPs that failed Illumina GoldenGate assay design.

Genetic association statistical analysis

Tests of Hardy–Weinberg equilibrium were calculated in the control group. A Bonferroni-corrected P-value threshold of 0.05/n was used to assess departure from Hardy–Weinberg equilibrium, where n was the number of SNPs successfully genotyped (n=5632). PLINK35 was used to calculate the Cochran–Armitage test for trend, the Breslow–Day test, the Cochran–Mantel–Haenszel tests for combined evidence of association across the two stages and the permutation test for correction for multiple testing (N=100000 permutations). Logistic regression was used to adjust for nongenetic factors such as age (Supplementary Table S3).


Conflict of interest

The authors declare no conflict of interest.



  1. The World Health Organization (2008). Hepatitis B. Fact sheet no. 204. Fact Sheet Revised August. World Health Organization: Geneva.
  2. Lavanchy D. Worldwide epidemiology of HBV infection, disease burden, and vaccine prevention. J Clin Virol 2005; 34 (Suppl 1) 1: S1–S3. | Article | PubMed
  3. Zuckerman JN. Protective efficacy, immunotherapeutic potential, and safety of hepatitis B vaccines. J Med Virol 2006; 78: 169–177. | Article | PubMed | ChemPort |
  4. Boland GJ, van Bommel T, Rulos-van den Berg A, van den Berg JP, van Loon AM, van Hattum J. The efficacy of a two-dose hepatitis B vaccination scheme. Adv Exp Med Biol 2003; 531: 185–190. | PubMed |
  5. Hollinger FB. Factors influencing the immune response to hepatitis B vaccine, booster dose guidelines, and vaccine protocol recommendations. Am J Med 1989; 87: 36S–40S. | Article | PubMed | ChemPort |
  6. Alper CA, Kruskall MS, Marcus-Bagley D, Craven DE, Katz AJ, Brink SJ et al. Genetic prediction of nonresponse to hepatitis B vaccine. N Engl J Med 1989; 321: 708–712. | PubMed | ISI | ChemPort |
  7. Hohler T, Reuss E, Evers N, Dietrich E, Rittner C, Freitag CM et al. Differential genetic determination of immune responsiveness to hepatitis B surface antigen and to hepatitis A virus: a vaccination study in twins. Lancet 2002; 360: 991–995. | Article | PubMed
  8. Milich DR, Leroux-Roels GG. Immunogenetics of the response to HBsAg vaccination. Autoimmun Rev 2003; 2: 248–257. | Article | PubMed | ChemPort |
  9. Newport MJ, Goetghebuer T, Weiss HA, Whittle H, Siegrist CA, Marchant A. Genetic regulation of immune responses to vaccines in early life. Genes Immun 2004; 5: 122–129. | Article | PubMed | ChemPort |
  10. McDermott AB, Zuckerman JN, Sabin CA, Marsh SG, Madrigal JA. Contribution of human leukocyte antigens to the antibody response to hepatitis B vaccination. Tissue Antigens 1997; 50: 8–14. | Article | PubMed | ISI | ChemPort |
  11. Desombere I, Willems A, Leroux-Roels G. Response to hepatitis B vaccine: multiple HLA genes are involved. Tissue Antigens 1998; 51: 593–604. | Article | PubMed | ISI | ChemPort |
  12. Yucesoy B, Sleijffers A, Kashon M, Garssen J, de Gruijl FR, Boland GJ et al. IL-1beta gene polymorphisms influence hepatitis B vaccination. Vaccine 2002; 20: 3193–3196. | Article | PubMed | ChemPort |
  13. Frodsham AJ, Zhang L, Dumpis U, Taib NA, Best S, Durham A et al. Class II cytokine receptor gene cluster is a major locus for hepatitis B persistence. Proc Natl Acad Sci USA 2006; 103: 9148–9153. | Article | PubMed | ChemPort |
  14. Thomas HC, Foster GR, Sumiya M, McIntosh D, Jack DL, Turner MW et al. Mutation of gene of mannose-binding protein associated with chronic hepatitis B viral infection. Lancet 1996; 348: 1417–1419. | Article | PubMed | ISI | ChemPort |
  15. Hohler T, Wunschel M, Gerken G, Schneider PM, Meyer zum Buschenfelde KH, Rittner C. No association between mannose-binding lectin alleles and susceptibility to chronic hepatitis B virus infection in German patients. Exp Clin Immunogenet 1998; 15: 130–133. | Article | PubMed | ChemPort |
  16. Thio CL, Mosbruger T, Astemborski J, Greer S, Kirk GD, O′Brien SJ et al. Mannose binding lectin genotypes influence recovery from hepatitis B virus infection. J Virol 2005; 79: 9192–9196. | Article | PubMed | ChemPort |
  17. Hennig BJ, Fielding K, Broxholme J, Diatta M, Mendy M, Moore C et al. Host genetic factors and vaccine-induced immunity to hepatitis B virus infection. PLoS ONE 2008; 3: e1898. | Article | PubMed | ChemPort |
  18. Hinds DA, Stuve LL, Nilsen GB, Halperin E, Eskin E, Ballinger DG et al. Whole-genome patterns of common DNA variation in three human populations. Science 2005; 307: 1072–1079. | Article | PubMed | ISI | ChemPort |
  19. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JP et al. Genome-wide association studies for complex traits: consensus, uncertainty and challenges. Nat Rev Genet 2008; 9: 356–369. | Article | PubMed | ChemPort |
  20. Balding DJ. A tutorial on statistical methods for population association studies. Nat Rev Genet 2006; 7: 781–791. | Article | PubMed | ISI | ChemPort |
  21. Wang C, Tang J, Song W, Lobashevsky E, Wilson CM, Kaslow RA. HLA and cytokine gene polymorphisms are independently associated with responses to hepatitis B vaccination. Hepatology 2004; 39: 978–988. | Article | PubMed | ChemPort |
  22. Hohler T, Meyer CU, Notghi A, Stradmann-Bellinghausen B, Schneider PM, Starke R et al. The influence of major histocompatibility complex class II genes and T-cell Vbeta repertoire on response to immunization with HBsAg. Hum Immunol 1998; 59: 212–218. | Article | PubMed | ChemPort |
  23. Hsu HY, Chang MH, Ho HN, Hsieh RP, Lee SD, Chen DS et al. Association of HLA-DR14-DR52 with low responsiveness to hepatitis B vaccine in Chinese residents in Taiwan. Vaccine 1993; 11: 1437–1440. | Article | PubMed | ChemPort |
  24. Stenzel A, Lu T, Koch WA, Hampe J, Guenther SM, De La Vega FM et al. Patterns of linkage disequilibrium in the MHC region on human chromosome 6p. Hum Genet 2004; 114: 377–385. | Article | PubMed | ISI | ChemPort |
  25. Kamatani Y, Wattanapokayakit S, Ochi H, Kawaguchi T, Takahashi A, Hosono N et al. A genome-wide association study identifies variants in the HLA-DP locus associated with chronic hepatitis B in Asians. Nat Genet 2009; 41: 591–595. | Article | PubMed | ChemPort |
  26. Hu H, Wang B, Borde M, Nardone J, Maika S, Allred L et al. Foxp1 is an essential transcriptional regulator of B cell development. Nat Immunol 2006; 7: 819–826. | Article | PubMed | ISI | ChemPort |
  27. Cella M, Dohring C, Samaridis J, Dessing M, Brockhaus M, Lanzavecchia A et al. A novel inhibitory receptor (ILT3) expressed on monocytes, macrophages, and dendritic cells involved in antigen processing. J Exp Med 1997; 185: 1743–1751. | Article | PubMed | ISI | ChemPort |
  28. Colonna M, Nakajima H, Cella M. A family of inhibitory and activating Ig-like receptors that modulate function of lymphoid and myeloid cells. Semin Immunol 2000; 12: 121–127. | Article | PubMed | ISI | ChemPort |
  29. Daheshia M, Friend DS, Grusby MJ, Austen KF, Katz HR. Increased severity of local and systemic anaphylactic reactions in gp49B1-deficient mice. J Exp Med 2001; 194: 227–234. | Article | PubMed | ISI | ChemPort |
  30. Chang CC, Ciubotariu R, Manavalan JS, Yuan J, Colovai AI, Piazza F et al. Tolerization of dendritic cells by T(S) cells: the crucial role of inhibitory receptors ILT3 and ILT4. Nat Immunol 2002; 3: 237–243. | Article | PubMed | ISI | ChemPort |
  31. Bissell DM. Chronic liver injury, TGF-beta, and cancer. Exp Mol Med 2001; 33: 179–190. | PubMed | ISI | ChemPort |
  32. Hashimoto K, Mori N, Tamesa T, Okada T, Kawauchi S, Oga A et al. Analysis of DNA copy number aberrations in hepatitis C virus-associated hepatocellular carcinomas by conventional CGH and array CGH. Mod Pathol 2004; 17: 617–622. | Article | PubMed | ISI | ChemPort |
  33. Plenge RM, Seielstad M, Padyukov L, Lee AT, Remmers EF, Ding B et al. TRAF1-C5 as a risk locus for rheumatoid arthritis—a genomewide study. N Engl J Med 2007; 357: 1199–1209. | Article | PubMed | ChemPort |
  34. European Consensus Group on hepatitis B immunity. Are booster immunizations needed for lifelong hepatitis B immunity? Lancet 2000; 355: 561–565. | ISI |
  35. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575. | Article | PubMed | ISI | ChemPort |
  36. Patterson N, Price AL, Reich D. Population structure and eigenanalysis. PLoS Genet 2006; 2: e190. | Article | PubMed | ChemPort |
  37. Langmore JP. Rubicon Genomics Inc. Pharmacogenomics 2002; 3: 557–560. | Article | PubMed
  38. Elston RC, Lin D, Zheng G. Multistage sampling for genetic studies. Annu Rev Genomics Hum Genet 2007; 8: 327–342. | Article | PubMed | ChemPort |
  39. Skol AD, Scott LJ, Abecasis GR, Boehnke M. Optimal designs for two-stage genome-wide association studies. Genet Epidemiol 2007; 31: 776–788. | Article | PubMed
  40. Thomas PD, Campbell MJ, Kejariwal A, Mi H, Karlak B, Daverman R et al. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res 2003; 13: 2129–2141. | Article | PubMed | ISI | ChemPort |
  41. Oliphant A, Barker DL, Stuelpnagel JR, Chee MS. BeadArray technology: enabling an accurate, cost-effective approach to high-throughput genotyping. Biotechniques 2002 (Suppl 56–58, 60–61), PMID: 12083399.
  42. Skipper L, Li Y, Bonnard C, Pavanni R, Yih Y, Chua E et al. Comprehensive evaluation of common genetic variation within LRRK2 reveals evidence for association with sporadic Parkinson's disease. Hum Mol Genet 2005; 14: 3549–3556. | Article | PubMed | ChemPort |


We are grateful to our colleagues at the Genome Institute of Singapore, Heng Khai Koon and Jason Ong for genotyping, Boon Yeong Goh for sample management and Frans Verhoef for maintaining the GISSNP database. We also thank Dr Jan van Hattum, Jan and Annemarie van den Berg for their dedication and enthusiasm in arranging and performing all the vaccinations. Finally, we thank Dr Sangkot Marzuki, Dr Herawati Sudoyo and Dr David Mulyono of the Eijkman Institute for Molecular Biology in Jakarta, Indonesia, for their invaluable advice, insights and assistance with this project. This work was supported by funding from the Agency for Science & Technology and Research of Singapore (A*STAR) and the US Defense Advanced Research Projects Agency Contract Number W911QY-06-C-0085. FEMF was supported by a Dr Saal van Zwanenberg fellowship.

Supplementary Information accompanies the paper on Genes and Immunity website