On the challenges of the HapMap resource

The International HapMap Project provides a key resource of genotypic data on human lymphoblastoid cell lines derived from four major world populations of European, African, Chinese and Japanese ancestry for researchers to associate with various phenotypic data to find genes affecting health, disease and response to drugs. Recently, the HapMap resource has significantly benefited research areas such as gene expression variation studies. Besides some intrinsic limitations, there are a few challenges that should be considered in the next wave of research using this tremendous resource. We suggest that overcoming these challenges or considering the confounding variables in the interpretation of results can provide more insights into the current views of the human genome as well as complex traits such as drug response variation and susceptibility to common diseases.     

Performance of genotype imputations using data from the 1000 Genomes Project

Genotype imputations based on 1000 Genomes (1KG) Project data have the advantage of imputing many more SNPs than imputations based on HapMap data. It also provides an opportunity to discover associations with relatively rare variants. Recent investigations are increasingly using 1KG data for genotype imputations, but only limited evaluations of the performance of this approach are available. In this paper, we empirically evaluated imputation performance using 1KG data by comparing imputation results to those using the HapMap Phase II data that have been widely used. We used three reference panels: the CEU panel consisting of 120 haplotypes from HapMap II and 1KG data (June 2010 release) and the EUR panel consisting of 566 haplotypes also from 1KG data (August 2010 release). We used Illumina 324,607 autosomal SNPs genotyped in 501 individuals of European ancestry. Our most important finding was that both 1KG reference panels provided much higher imputation yield than the HapMap II panel. There were more than twice as many successfully imputed SNPs as there were using the HapMap II panel (6.7 million vs. 2.5 million). Our second most important finding was that accuracy using both 1KG panels was high and almost identical to accuracy using the HapMap II panel. Furthermore, after removing SNPs with MACH Rsq <0.3, accuracy for both rare and low frequency SNPs was very high and almost identical to accuracy for common SNPs. We found that imputation using the 1KG-EUR panel had advantages in successfully imputing rare, low frequency and common variants. Our findings suggest that 1KG-based imputation can increase the opportunity to discover significant associations for SNPs across the allele frequency spectrum. Because the 1KG Project is still underway, we expect that later versions will provide even better imputation performance.     

Comparative study of the linkage disequilibrium of an ENCODE region, chromosome 7p15, in Korean, Japanese, and Han Chinese samples

The extent and pattern of linkage disequilibrium (LD) in the human genome provide important information for disease gene mapping. Previous studies have shown that LDs vary depending on chromosomal regions and populations. As the Asian samples of the International HapMap Project consisted of Japanese and Chinese populations, it was of interest whether we could use the HapMap data as a reference to carry out association studies of common complex diseases in a closely related population, such as Koreans. We have compared the LD and recombination patterns defined by single-nucleotide polymorphisms (SNPs) in ENCODE region ENm010, chromosome 7p15.2, in Korean, Japanese, and Chinese samples and further tested the robustness of tagSNPs among the Asian samples. We genotyped 792 SNPs in 500 kb (chromosome 7: 26699793-27199792, NCBI build 34) from 90 unrelated Koreans by fluorescence polarization detection and compared the data with Asian data from the HapMap project. Despite some differences in the position of high LD region boundaries, the overall patterns of LD were remarkably similar across the three samples, reflecting strong genetic affinities among them. Furthermore, the haplotype tag SNP transferability across the three samples was greater than 90%. Our results support the initial suggestion that the populations genotyped in the HapMap project might serve as reference populations for the selection of tagSNPs in association studies.     

Investigation of inversion polymorphisms in the human genome using principal components analysis

Despite the significant advances made over the last few years in mapping inversions with the advent of paired-end sequencing approaches, our understanding of the prevalence and spectrum of inversions in the human genome has lagged behind other types of structural variants, mainly due to the lack of a cost-efficient method applicable to large-scale samples. We propose a novel method based on principal components analysis (PCA) to characterize inversion polymorphisms using high-density SNP genotype data. Our method applies to non-recurrent inversions for which recombination between the inverted and non-inverted segments in inversion heterozygotes is suppressed due to the loss of unbalanced gametes. Inside such an inversion region, an effect similar to population substructure is thus created: two distinct "populations" of inversion homozygotes of different orientations and their 1:1 admixture, namely the inversion heterozygotes. This kind of substructure can be readily detected by performing PCA locally in the inversion regions. Using simulations, we demonstrated that the proposed method can be used to detect and genotype inversion polymorphisms using unphased genotype data. We applied our method to the phase III HapMap data and inferred the inversion genotypes of known inversion polymorphisms at 8p23.1 and 17q21.31. These inversion genotypes were validated by comparing with literature results and by checking Mendelian consistency using the family data whenever available. Based on the PCA-approach, we also performed a preliminary genome-wide scan for inversions using the HapMap data, which resulted in 2040 candidate inversions, 169 of which overlapped with previously reported inversions. Our method can be readily applied to the abundant SNP data, and is expected to play an important role in developing human genome maps of inversions and exploring associations between inversions and susceptibility of diseases.     

Targeted sequencing of the human X chromosome exome

We used a RainDance Technologies (RDT) expanded content library to enrich the human X chromosome exome (2.5 Mb) from 26 male samples followed by Illumina sequencing. Our multiplex primer library covered 98.05% of the human X chromosome exome in a single tube with 11,845 different PCR amplicons. Illumina sequencing of 24 male samples showed coverage for 97% of the targeted sequences. Sequence from 2 HapMap samples confirmed missing data rates of 2–3% at sites successfully typed by the HapMap project, with an accuracy of at least ~ 99.5% as compared to reported HapMap genotypes. Our demonstration that a RDT expanded content library can efficiently enrich and enable the routine sequencing of the human X chromosome exome suggests a wide variety of potential research and clinical applications for this platform.     

Genetic Variation in TLR Genes in Ugandan and South African Populations and Comparison with HapMap Data

Genetic epidemiological studies of complex diseases often rely on data from the International HapMap Consortium for identification of single nucleotide polymorphisms (SNPs), particularly those that tag haplotypes. However, little is known about the relevance of the African populations used to collect HapMap data for study populations conducted elsewhere in Africa. Toll-like receptor (TLR) genes play a key role in susceptibility to various infectious diseases, including tuberculosis. We conducted full-exon sequencing in samples obtained from Uganda (n = 48) and South Africa (n = 48), in four genes in the TLR pathway: TLR2, TLR4, TLR6, and TIRAP. We identified one novel TIRAP SNP (with minor allele frequency [MAF] 3.2%) and a novel TLR6 SNP (MAF 8%) in the Ugandan population, and a TLR6 SNP that is unique to the South African population (MAF 14%). These SNPs were also not present in the 1000 Genomes data. Genotype and haplotype frequencies and linkage disequilibrium patterns in Uganda and South Africa were similar to African populations in the HapMap datasets. Multidimensional scaling analysis of polymorphisms in all four genes suggested broad overlap of all of the examined African populations. Based on these data, we propose that there is enough similarity among African populations represented in the HapMap database to justify initial SNP selection for genetic epidemiological studies in Uganda and South Africa. We also discovered three novel polymorphisms that appear to be population-specific and would only be detected by sequencing efforts.     

Comparison of linkage disequilibrium patterns between the HapMap CEPH samples and a family-based cohort of Northern European descent

The International HapMap Consortium has determined the linkage disequilibrium (LD) patterns of four major human populations. The aim of our investigation was to compare the LD patterns of the HapMap CEPH (Centre d'Etude du Polymorphisme Humain) samples with a family-based cohort of similar ancestry to determine its usefulness as a reference population for disease association studies. We examined four genomic regions on chromosomes 7q, 12p, and 14q totaling 14.3 Mb, initially identified in our linkage study of obesity and the metabolic syndrome. Near identical patterns of LD were detected in both populations. Furthermore, tagSNPs selected based on the HapMap CEPH cohort data capture over 98% of the variants at an r2 > 0.8 in the disease cohort. This confirms the usefulness of the CEPH cohort of the HapMap as a reference sample for further investigations into the genomic variation of populations of Northern European descent.     

Identification of the ancestral killer immunoglobulin-like receptor gene in primates

Background

The recent advancement in human genome sequencing and genotyping has revealed millions of single nucleotide polymorphisms (SNP) which determine the variation among human beings. One of the particular important projects is The International HapMap Project which provides the catalogue of human genetic variation for disease association studies. In this paper, we analyzed the genotype data in HapMap project by using National Institute of Environmental Health Sciences Environmental Genome Project (NIEHS EGP) SNPs. We first determine whether the HapMap data are transferable to the NIEHS data. Then, we study how well the HapMap SNPs capture the untyped SNPs in the region. Finally, we provide general guidelines for determining whether the SNPs chosen from HapMap may be able to capture most of the untyped SNPs.

Results

Our analysis shows that HapMap data are not robust enough to capture the untyped variants for most of the human genes. The performance of SNPs for European and Asian samples are marginal in capturing the untyped variants, i.e. approximately 55%. Expectedly, the SNPs from HapMap YRI panel can only capture approximately 30% of the variants. Although the overall performance is low, however, the SNPs for some genes perform very well and are able to capture most of the variants along the gene. This is observed in the European and Asian panel, but not in African panel. Through observation, we concluded that in order to have a well covered SNPs reference panel, the SNPs density and the association among reference SNPs are important to estimate the robustness of the chosen SNPs.

Conclusion

We have analyzed the coverage of HapMap SNPs using NIEHS EGP data. The results show that HapMap SNPs are transferable to the NIEHS SNPs. However, HapMap SNPs cannot capture some of the untyped SNPs and therefore resequencing may be needed to uncover more SNPs in the missing region.     

Arrayed primer extension in the "array of arrays" format: a rational approach for microarray-based SNP genotyping

This study provides a new version of the arrayed primer extension (APEX) protocol adapted to the 'array of arrays' platform using an instrumental setup for microarray processing not previously described. The primary aim of the study is to implement a system for rational cost-efficient genotyping where multiple singlenucleotide polymorphisms (SNPs) and individuals are genotyped on each microarray slide. Genotyping results are collected across 185 healthy Danish subjects and 76 SNPs on chromosome 3q13.31, because linkage to atopic disease phenotypes have been suggested in the Danish population. Linkage disequilibrium (LD) results from the experimental data are used in a novel comparison to baseline data defined by the international HapMap SNP database. Comparison on the LD results reveals a strong linear correlation irrespective of LD measure considered: R2 (D') = 0.73 and R2(r2) = 0.54. In conclusion, our results show that this setup is strong enough to support high-throughput genotyping, and these observations support that the HapMap genotype resource is important for defining SNP panels aimed at gene mapping in local subpopulations from Europe.     

FstSNP-HapMap3: a database of SNPs with high population differentiation for HapMap3

The International HapMap Project has recently made available genotypes and frequency data for phase 3 (NCBI build 36, dbSNPb129) of the HapMap providing an enriched genotype dataset for approximately 1.6 million single nucleotide polymorphisms (SNPs) from 1,115 individuals with ancestry from parts of Africa, Asia, Europe, North America and Mexico. In the present study, we aim to facilitate pharmacogenetics studies by providing a database of SNPs with high population differentiation through a genomewide test on allele frequency variation among 11 HapMap3 samples. Common SNPs with minor allele frequency greater than 5¢ from each of 11 HapMap3 samples were included in the present analysis. The population differentiation is measured in terms of fixation index (Fst), and the SNPs with Fst values over 0.5 were defined as highly differentiated SNPs. Our tests were carried out between all pairs of the 11 HapMap3 samples or among subgroups with the same continental ancestries. Altogether we carried out 64 genomewide Fst tests and identified 28,215 highly differentiated SNPs for 49 different combinations of HapMap3 samples in the current database.     

Large scale association analysis identifies three susceptibility loci for coronary artery disease

Genome wide association studies (GWAS) and their replications that have associated DNA variants with myocardial infarction (MI) and/or coronary artery disease (CAD) are predominantly based on populations of European or Eastern Asian descent. Replication of the most significantly associated polymorphisms in multiple populations with distinctive genetic backgrounds and lifestyles is crucial to the understanding of the pathophysiology of a multifactorial disease like CAD. We have used our Lebanese cohort to perform a replication study of nine previously identified CAD/MI susceptibility loci (LTA, CDKN2A-CDKN2B, CELSR2-PSRC1-SORT1, CXCL12, MTHFD1L, WDR12, PCSK9, SH2B3, and SLC22A3), and 88 genes in related phenotypes. The study was conducted on 2,002 patients with detailed demographic, clinical characteristics, and cardiac catheterization results. One marker, rs6922269, in MTHFD1L was significantly protective against MI (OR=0.68, p=0.0035), while the variant rs4977574 in CDKN2A-CDKN2B was significantly associated with MI (OR=1.33, p=0.0086). Associations were detected after adjustment for family history of CAD, gender, hypertension, hyperlipidemia, diabetes, and smoking. The parallel study of 88 previously published genes in related phenotypes encompassed 20,225 markers, three quarters of which with imputed genotypes The study was based on our genome-wide genotype data set, with imputation across the whole genome to HapMap II release 22 using HapMap CEU population as a reference. Analysis was conducted on both the genotyped and imputed variants in the 88 regions covering selected genes. This approach replicated HNRNPA3P1-CXCL12 association with CAD and identified new significant associations of CDKAL1, ST6GAL1, and PTPRD with CAD. Our study provides evidence for the importance of the multifactorial aspect of CAD/MI and describes genes predisposing to their etiology.     

Fine-scale mapping of disease susceptibility locus with Bayesian partition model

The causal relationship between genes and diseases has been investigated with the development of DNA sequence. Polymorphisms incorporated in the HapMap Project have enabled fine mapping with linkage disequilibrium (LD) and prior clustering of the haplotypes on the basis of a similarity measure has often been performed in an attempt to capture coalescent events because they can reduce the amount of computation. However an inappropriate choice of similarity measure can lead to wrong conclusions and we propose a new haplotype-based clustering algorithm for fine-scale mapping by using a Bayesian partition model. To handle phase-unknown genotypes, we propose a new algorithm based on a Metropolized Gibbs sampler and it is implemented in C++. Our simulation studies found that the proposed method improves the accuracy of the estimator for the disease susceptibility locus. We illustrated the practical implication of the new analysis method by an application to fine-scale mapping of CYP2D6 in drug metabolism.     

Crohn's disease risk alleles on the NOD2 locus have been maintained by natural selection on standing variation

Risk alleles for complex diseases are widely spread throughout human populations. However, little is known about the geographic distribution and frequencies of risk alleles, which may contribute to differences in disease susceptibility and prevalence among populations. Here, we focus on Crohn's disease (CD) as a model for the evolutionary study of complex disease alleles. Recent genome-wide association studies and classical linkage analyses have identified more than 70 susceptible genomic regions for CD in Europeans, but only a few have been confirmed in non-European populations. Our analysis of eight European-specific susceptibility genes using HapMap data shows that at the NOD2 locus the CD-risk alleles are linked with a haplotype specific to CEU at a frequency that is significantly higher compared with the entire genome. We subsequently examined nine global populations and found that the CD-risk alleles spread through hitchhiking with a high-frequency haplotype (H1) exclusive to Europeans. To examine the neutrality of NOD2, we performed phylogenetic network analyses, coalescent simulation, protein structural prediction, characterization of mutation patterns, and estimations of population growth and time to most recent common ancestor (TMRCA). We found that while H1 was significantly prevalent in European populations, the H1 TMRCA predated human migration out of Africa. H1 is likely to have undergone negative selection because 1) the root of H1 genealogy is defined by a preexisting amino acid substitution that causes serious conformational changes to the NOD2 protein, 2) the haplotype has almost become extinct in Africa, and 3) the haplotype has not been affected by the recent European expansion reflected in the other haplotypes. Nevertheless, H1 has survived in European populations, suggesting that the haplotype is advantageous to this group. We propose that several CD-risk alleles, which destabilize and disrupt the NOD2 protein, have been maintained by natural selection on standing variation because the deleterious haplotype of NOD2 is advantageous in diploid individuals due to heterozygote advantage and/or intergenic interactions.     

Simulating association studies: a data-based resampling method for candidate regions or whole genome scans

MOTIVATION: Reductions in genotyping costs have heightened interest in performing whole genome association scans and in the fine mapping of candidate regions. Improvements in study design and analytic techniques will require the simulation of datasets with realistic patterns of linkage disequilibrium and allele frequencies for typed SNPs. METHODS: We describe a general approach to simulate genotyped datasets for standard case-control or affected child trio data, by resampling from existing phased datasets. The approach allows for considerable flexibility in disease models, potentially involving a large number of interacting loci. The method is most applicable for diseases caused by common variants that have not been under strong selection, a class specifically targeted by the International HapMap project. RESULTS: Using the three population Phase I/II HapMap data as a testbed for our approach, we have implemented the approach in HAP-SAMPLE, a web-based simulation tool.     

Evaluating HapMap SNP data transferability in a large-scale genotyping project involving 175 cancer-associated genes

One of the many potential uses of the HapMap project is its application to the investigation of complex disease aetiology among a wide range of populations. This study aims to assess the transferability of HapMap SNP data to the Spanish population in the context of cancer research. We have carried out a genotyping study in Spanish subjects involving 175 candidate cancer genes using an indirect gene-based approach and compared results with those for HapMap CEU subjects. Allele frequencies were very consistent between the two samples, with a high positive correlation (R) of 0.91 (P<<1×10⁻⁶). Linkage disequilibrium patterns and block structures across each gene were also very similar, with disequilibrium coefficient (r ²) highly correlated (R=0.95, P<<1×10⁻⁶). We found that of the 21 genes that contained at least one block larger than 60 kb, nine (ATM, ATR, BRCA1, ERCC6, FANCC, RAD17, RAD50, RAD54B and XRCC4) belonged to the GO category “DNA repair”. Haplotype frequencies per gene were also highly correlated (mean R=0.93), as was haplotype diversity (R=0.91, P<<1×10⁻⁶). “Yin yang” haplotypes were observed for 43% of the genes analysed and 18% of those were identical to the ancestral haplotype (identified in Chimpazee). Finally, the portability of tagSNPs identified in the HapMap CEU data using pairwise r ² thresholds of 0.8 and 0.5 was assessed by applying these to the Spanish and current HapMap data for 66 genes. In general, the HapMap tagSNPs performed very well. Our results show generally high concordance with HapMap data in allele frequencies and haplotype distributions and confirm the applicability of HapMap SNP data to the study of complex diseases among the Spanish population. Electronic Supplementary Material Supplementary material is available for this article at and is accessible for authorized users.     

GWAsimulator: a rapid whole-genome simulation program

SUMMARY: GWAsimulator implements a rapid moving-window algorithm to simulate genotype data for case-control or population samples from genomic SNP chips. For case-control data, the program generates cases and controls according to a user-specified multi-locus disease model, and can simulate specific regions if desired. The program uses phased genotype data as input and has the flexibility of simulating genotypes for different populations and different genomic SNP chips. When the HapMap phased data are used, the simulated data have similar local LD patterns as the HapMap data. As genome-wide association (GWA) studies become increasingly popular and new GWA data analysis methods are being developed, we anticipate that GWAsimulator will be an important tool for evaluating performance of new GWA analysis methods. AVAILABILITY: The C++ source code, executables for Linux, Windows and MacOS, manual, example data sets and analysis program are available at http://biostat.mc.vanderbilt.edu/GWAsimulator     

Combinatorial effects of four histone modifications in transcription and differentiation

Linkage disequilibrium (LD) has received much attention recently because of its value in localizing disease-causing genes. Due to the extensive LD between neighboring loci in the human genome, it is believed that a subset of the single nucleotide polymorphisms in a region (tagSNPs) can be selected to capture most of the remaining SNP variants. In this study, we examined LD patterns and HapMap tagSNP transferability in more than 300 individuals. A South Indian sample and an African Mbuti Pygmy population sample were included to evaluate the performance of HapMap tagSNPs in geographically distinct and genetically isolated populations. Our results show that HapMap tagSNPs selected with r(2) >= 0.8 can capture more than 85% of the SNPs in populations that are from the same continental group. Combined tagSNPs from HapMap CEU and CHB+JPT serve as the best reference for the Indian sample. The HapMap YRI are a sufficient reference for tagSNP selection in the Pygmy sample. In addition to our findings, we reviewed over 25 recent studies of tagSNP transferability and propose a general guideline for selecting tagSNPs from HapMap populations.     

Assessing Accuracy of Genotype Imputation in American Indians

Background

Genotype imputation is commonly used in genetic association studies to test untyped variants using information on linkage disequilibrium (LD) with typed markers. Imputing genotypes requires a suitable reference population in which the LD pattern is known, most often one selected from HapMap. However, some populations, such as American Indians, are not represented in HapMap. In the present study, we assessed accuracy of imputation using HapMap reference populations in a genome-wide association study in Pima Indians.

Results

Data from six randomly selected chromosomes were used. Genotypes in the study population were masked (either 1% or 20% of SNPs available for a given chromosome). The masked genotypes were then imputed using the software Markov Chain Haplotyping Algorithm. Using four HapMap reference populations, average genotype error rates ranged from 7.86% for Mexican Americans to 22.30% for Yoruba. In contrast, use of the original Pima Indian data as a reference resulted in an average error rate of 1.73%.

Conclusions

Our results suggest that the use of HapMap reference populations results in substantial inaccuracy in the imputation of genotypes in American Indians. A possible solution would be to densely genotype or sequence a reference American Indian population.     

A comparison of phasing algorithms for trios and unrelated individuals

Knowledge of haplotype phase is valuable for many analysis methods in the study of disease, population, and evolutionary genetics. Considerable research effort has been devoted to the development of statistical and computational methods that infer haplotype phase from genotype data. Although a substantial number of such methods have been developed, they have focused principally on inference from unrelated individuals, and comparisons between methods have been rather limited. Here, we describe the extension of five leading algorithms for phase inference for handling father-mother-child trios. We performed a comprehensive assessment of the methods applied to both trios and to unrelated individuals, with a focus on genomic-scale problems, using both simulated data and data from the HapMap project. The most accurate algorithm was PHASE (v2.1). For this method, the percentages of genotypes whose phase was incorrectly inferred were 0.12%, 0.05%, and 0.16% for trios from simulated data, HapMap Centre d'Etude du Polymorphisme Humain (CEPH) trios, and HapMap Yoruban trios, respectively, and 5.2% and 5.9% for unrelated individuals in simulated data and the HapMap CEPH data, respectively. The other methods considered in this work had comparable but slightly worse error rates. The error rates for trios are similar to the levels of genotyping error and missing data expected. We thus conclude that all the methods considered will provide highly accurate estimates of haplotypes when applied to trio data sets. Running times differ substantially between methods. Although it is one of the slowest methods, PHASE (v2.1) was used to infer haplotypes for the 1 million-SNP HapMap data set. Finally, we evaluated methods of estimating the value of r(2) between a pair of SNPs and concluded that all methods estimated r(2) well when the estimated value was >or=0.8.