De novo assembly and genotyping of variants using colored de Bruijn graphs

Detecting genetic variants that are highly divergent from a reference sequence remains a major challenge in genome sequencing. We introduce de novo assembly algorithms using colored de Bruijn graphs for detecting and genotyping simple and complex genetic variants in an individual or population. We provide an efficient software implementation, Cortex, the first de novo assembler capable of assembling multiple eukaryotic genomes simultaneously. Four applications of Cortex are presented. First, we detect and validate both simple and complex structural variations in a high-coverage human genome. Second, we identify more than 3 Mb of sequence absent from the human reference genome, in pooled low-coverage population sequence data from the 1000 Genomes Project. Third, we show how population information from ten chimpanzees enables accurate variant calls without a reference sequence. Last, we estimate classical human leukocyte antigen (HLA) genotypes at HLA-B, the most variable gene in the human genome.     

A predominant role for the HLA class II region in the association of the MHC region with multiple sclerosis

Genetic susceptibility to multiple sclerosis is associated with genes of the major histocompatibility complex (MHC), particularly HLA-DRB1 and HLA-DQB1 (ref. 1). Both locus and allelic heterogeneity have been reported in this genomic region. To clarify whether HLA-DRB1 itself, nearby genes in the region encoding the MHC or combinations of these loci underlie susceptibility to multiple sclerosis, we genotyped 1,185 Canadian and Finnish families with multiple sclerosis (n = 4,203 individuals) with a high-density SNP panel spanning the genes encoding the MHC and flanking genomic regions. Strong associations in Canadian and Finnish samples were observed with blocks in the HLA class II genomic region (P < 4.9 x 10(-13) and P < 2.0 x 10(-16), respectively), but the strongest association was with HLA-DRB1 (P < 4.4 x 10(-17)). Conditioning on either HLA-DRB1 or the most significant HLA class II haplotype block found no additional block or SNP association independent of the HLA class II genomic region. This study therefore indicates that MHC-associated susceptibility to multiple sclerosis is determined by HLA class II alleles, their interactions and closely neighboring variants.     

Epistatic interaction between KIR3DS1 and HLA-B delays the progression to AIDS

Natural killer (NK) cells provide defense in the early stages of the innate immune response against viral infections by producing cytokines and causing cytotoxicity. The killer immunoglobulin-like receptors (KIRs) on NK cells regulate the inhibition and activation of NK-cell responses through recognition of human leukocyte antigen (HLA) class I molecules on target cells KIR and HLA loci are both highly polymorphic, and some HLA class I products bind and trigger cell-surface receptors specified by KIR genes. Here we report that the activating KIR allele KIR3DS1, in combination with HLA-B alleles that encode molecules with isoleucine at position 80 (HLA-B Bw4-80Ile), is associated with delayed progression to AIDS in individuals infected with human immunodeficiency virus type 1 (HIV-1). In the absence of KIR3DS1, the HLA-B Bw4-80Ile allele was not associated with any of the AIDS outcomes measured. By contrast, in the absence of HLA-B Bw4-80Ile alleles, KIR3DS1 was significantly associated with more rapid progression to AIDS. These observations are strongly suggestive of a model involving an epistatic interaction between the two loci. The strongest synergistic effect of these loci was on progression to depletion of CD4(+) T cells, which suggests that a protective response of NK cells involving KIR3DS1 and its HLA class I ligands begins soon after HIV-1 infection.     

A high-resolution HLA and SNP haplotype map for disease association studies in the extended human MHC

The proteins encoded by the classical HLA class I and class II genes in the major histocompatibility complex (MHC) are highly polymorphic and are essential in self versus non-self immune recognition. HLA variation is a crucial determinant of transplant rejection and susceptibility to a large number of infectious and autoimmune diseases. Yet identification of causal variants is problematic owing to linkage disequilibrium that extends across multiple HLA and non-HLA genes in the MHC. We therefore set out to characterize the linkage disequilibrium patterns between the highly polymorphic HLA genes and background variation by typing the classical HLA genes and >7,500 common SNPs and deletion-insertion polymorphisms across four population samples. The analysis provides informative tag SNPs that capture much of the common variation in the MHC region and that could be used in disease association studies, and it provides new insight into the evolutionary dynamics and ancestral origins of the HLA loci and their haplotypes.     

Global diversity and evidence for coevolution of KIR and HLA

The killer immunoglobulin-like receptor (KIR) gene cluster shows extensive genetic diversity, as do the HLA class I loci, which encode ligands for KIR molecules. We genotyped 1,642 individuals from 30 geographically distinct populations to examine population-level evidence for coevolution of these two functionally related but unlinked gene clusters. We observed strong negative correlations between the presence of activating KIR genes and their corresponding HLA ligand groups across populations, especially KIR3DS1 and its putative HLA-B Bw4-80I ligands (r = -0.66, P = 0.038). In contrast, we observed weak positive relationships between the various inhibitory KIR genes and their ligands. We observed a negative correlation between distance from East Africa and frequency of activating KIR genes and their corresponding ligands, suggesting a balance between selection on HLA and KIR loci. Most KIR-HLA genetic association studies indicate a primary influence of activating KIR-HLA genotypes in disease risk; concomitantly, activating receptor-ligand pairs in this study show the strongest signature of coevolution of these two complex genetic systems as compared with inhibitory receptor-ligand pairs.     

Linkage disequilibrium of a type 1 diabetes susceptibility locus with a regulatory IL12B allele

Type 1 diabetes (T1D; or insulin-dependent diabetes mellitus, IDDM) is an autoimmune disease with both genetic and environmental components. In addition to the human leukocyte antigen (HLA) complex, the single major genetic contributor of susceptibility, an unknown number of other unidentified genes are required to mediate disease. Although many loci conferring susceptibility to T1D have been mapped, their identification has proven problematic due to the complex nature of this disease. Our strategy for finding T1D susceptibility genes has been to test for human homologues of loci implicated in diabetes-prone NOD (non-obese diabetic) mice, together with application of biologically relevant stratification methods. We report here a new susceptibility locus, IDDM18, located near the interleukin-12 (IL-12)p40 gene, IL12B. Significant bias in transmission of IL12B alleles was observed in affected sibpairs and was confirmed in an independent cohort of simplex families. A single base change in the 3' UTR showed strong linkage disequilibrium with the T1D susceptibility locus. The IL12B 3' UTR alleles showed different levels of expression in cell lines. Variation in IL-12p40 production may influence T-cell responses crucial for either mediating or protecting against this and other autoimmune diseases.     

Heterogeneous mutation processes in human microsatellite DNA sequences

Although microsatellite polymorphisms are one of the most commonly used tools in genetic analyses, it remains to be understood how microsatellite DNA has evolved as a ubiquitous and highly abundant class of repetitive sequences in eukaryotic genomes. On the basis of analyses of spontaneous human microsatellite mutations of germline origin, I show here that different mutation biases underlie the evolution of microsatellite repeats. The within-locus mutation rate increases with allele length, but is not affected by the size difference between an individual's two alleles (allele span). Within loci, long alleles tend to mutate to shorter lengths, thereby acting to prevent infinite growth. Expansions are more common than contractions among dinucleotide repeats, whereas no such trend is evident among tetranucleotide repeats. This observation is consistent with the longer repeat lengths and higher frequency of di- compared with tetranucleotide repeats. An excess of paternally transmitted mutations (male-to-female ratio of 4.9) supports a male-biased mutation rate in the human genome.     

Nuclear genetic control of mitochondrial DNA segregation

Mammalian mitochondrial DNA (mtDNA) is a high copy-number, maternally inherited genome that codes for a small number of essential proteins involved in oxidative phosphorylation. Mutations in mtDNA are responsible for a broad spectrum of clinical disorders. The segregation pattern of pathogenic mtDNA mutants is an important determinant of the nature and severity of mitochondrial disease, but it varies with the specific mutation, cell type and nuclear background and generally does not correlate well with mitochondrial dysfunction. To identify nuclear genes that modify the segregation behavior of mtDNA, we used a heteroplasmic mouse model derived from two inbred strains (BALB/c and NZB; ref. 12), in which we had previously demonstrated tissue-specific and age-dependent directional selection for different mtDNA genotypes in the same mouse. Here we show that this phenotype segregates in F2 mice from a genetic cross (BALB/c x CAST/Ei) and that it maps to at least three quantitative-trait loci (QTLs). Genome-wide scans showed linkage of the trait to loci on Chromosomes 2, 5 and 6, accounting for 16-35% of the variance in the trait, depending on the tissue and age of the mouse. This is the first genetic evidence for nuclear control of mammalian mtDNA segregation.     

Spatial proximity of translocation-prone gene loci in human lymphomas

Cancer cells frequently have disease-specific chromosome rearrangements. It is poorly understood why translocations between chromosomes recur at specific breakpoints in the genome. Here we provide evidence that higher-order spatial genome organization is a contributing factor in the formation of recurrent translocations. We show that MYC, BCL and immunoglobulin loci, which are recurrently translocated in various B-cell lymphomas, are preferentially positioned in close spatial proximity relative to each other in normal B cells. Loci in spatial proximity are non-randomly positioned towards the nuclear interior in normal B cells. This locus proximity is the consequence of higher-order genome structure rather than a property of individual genes. Our results suggest that the formation of specific translocations in human lymphomas, and perhaps other tissues, is determined in part by higher-order spatial organization of the genome.     

Intensely punctate meiotic recombination in the class II region of the major histocompatibility complex

There is considerable interest in understanding patterns of linkage disequilibrium (LD) in the human genome, to aid investigations of human evolution and facilitate association studies in complex disease. The relative influences of meiotic crossover distribution and population history on LD remain unclear, however. In particular, it is uncertain to what extent crossovers are clustered into 'hot spots, that might influence LD patterns. As a first step to investigating the relationship between LD and recombination, we have analyzed a 216-kb segment of the class II region of the major histocompatibility complex (MHC) already characterized for familial crossovers. High-resolution LD analysis shows the existence of extended domains of strong association interrupted by patchwork areas of LD breakdown. Sperm typing shows that these areas correspond precisely to meiotic crossover hot spots. All six hot spots defined share a remarkably similar symmetrical morphology but vary considerably in intensity, and are not obviously associated with any primary DNA sequence determinants of hot-spot activity. These hot spots occur in clusters and together account for almost all crossovers in this region of the MHC. These data show that, within the MHC at least, crossovers are far from randomly distributed at the molecular level and that recombination hot spots can profoundly affect LD patterns.     

Identification of ten loci associated with height highlights new biological pathways in human growth

Height is a classic polygenic trait, reflecting the combined influence of multiple as-yet-undiscovered genetic factors. We carried out a meta-analysis of genome-wide association study data of height from 15,821 individuals at 2.2 million SNPs, and followed up the strongest findings in >10,000 subjects. Ten newly identified and two previously reported loci were strongly associated with variation in height (P values from 4 x 10(-7) to 8 x 10(-22)). Together, these 12 loci account for approximately 2% of the population variation in height. Individuals with < or =8 height-increasing alleles and > or =16 height-increasing alleles differ in height by approximately 3.5 cm. The newly identified loci, along with several additional loci with strongly suggestive associations, encompass both strong biological candidates and unexpected genes, and highlight several pathways (let-7 targets, chromatin remodeling proteins and Hedgehog signaling) as important regulators of human stature. These results expand the picture of the biological regulation of human height and of the genetic architecture of this classical complex trait.     

The extent of linkage disequilibrium in Arabidopsis thaliana.

Linkage disequilibrium (LD), the nonrandom occurrence of alleles in haplotypes, has long been of interest to population geneticists. Recently, the rapidly increasing availability of genomic polymorphism data has fueled interest in LD as a tool for fine-scale mapping, in particular for human disease loci. The chromosomal extent of LD is crucial in this context, because it determines how dense a map must be for associations to be detected and, conversely, limits how finely loci may be mapped. Arabidopsis thaliana is expected to harbor unusually extensive LD because of its high degree of selfing. Several polymorphism studies have found very strong LD within individual loci, but also evidence of some recombination. Here we investigate the pattern of LD on a genomic scale and show that in global samples, LD decays within approximately 1 cM, or 250 kb. We also show that LD in local populations may be much stronger than that of global populations, presumably as a result of founder events. The combination of a relatively high level of polymorphism and extensive haplotype structure bodes well for developing a genome-wide LD map in A. thaliana.     

Genome-wide association study identifies susceptibility loci for dengue shock syndrome at MICB and PLCE1

Hypovolemic shock (dengue shock syndrome (DSS)) is the most common life-threatening complication of dengue. We conducted a genome-wide association study of 2,008 pediatric cases treated for DSS and 2,018 controls from Vietnam. Replication of the most significantly associated markers was carried out in an independent Vietnamese sample of 1,737 cases and 2,934 controls. SNPs at two loci showed genome-wide significant association with DSS. We identified a susceptibility locus at MICB (major histocompatibility complex (MHC) class I polypeptide-related sequence B), which was within the broad MHC region on chromosome 6 but outside the class I and class II HLA loci (rs3132468, P(meta) = 4.41 × 10(-11), per-allele odds ratio (OR) = 1.34 (95% confidence interval: 1.23-1.46)). We identified associated variants within PLCE1 (phospholipase C, epsilon 1) on chromosome 10 (rs3765524, P(meta) = 3.08 × 10(-10), per-allele OR = 0.80 (95% confidence interval: 0.75-0.86)). We identify two loci associated with susceptibility to DSS in people with dengue, suggesting possible mechanisms for this severe complication of dengue.     

Evidence of widespread selection on standing variation in Europe at height-associated SNPs

Strong signatures of positive selection at newly arising genetic variants are well documented in humans, but this form of selection may not be widespread in recent human evolution. Because many human traits are highly polygenic and partly determined by common, ancient genetic variation, an alternative model for rapid genetic adaptation has been proposed: weak selection acting on many pre-existing (standing) genetic variants, or polygenic adaptation. By studying height, a classic polygenic trait, we demonstrate the first human signature of widespread selection on standing variation. We show that frequencies of alleles associated with increased height, both at known loci and genome wide, are systematically elevated in Northern Europeans compared with Southern Europeans (P < 4.3 × 10(-4)). This pattern mirrors intra-European height differences and is not confounded by ancestry or other ascertainment biases. The systematic frequency differences are consistent with the presence of widespread weak selection (selection coefficients ～10(-3)-10(-5) per allele) rather than genetic drift alone (P < 10(-15)).     

Genetics of gene expression in primary immune cells identifies cell type-specific master regulators and roles of HLA alleles

Trans-acting genetic variants have a substantial, albeit poorly characterized, role in the heritable determination of gene expression. Using paired purified primary monocytes and B cells, we identify new predominantly cell type-specific cis and trans expression quantitative trait loci (eQTLs), including multi-locus trans associations to LYZ and KLF4 in monocytes and B cells, respectively. Additionally, we observe a B cell-specific trans association of rs11171739 at 12q13.2, a known autoimmune disease locus, with IP6K2 (P = 5.8 × 10(-15)), PRIC285 (P = 3.0 × 10(-10)) and an upstream region of CDKN1A (P = 2 × 10(-52)), suggesting roles for cell cycle regulation and peroxisome proliferator-activated receptor γ (PPARγ) signaling in autoimmune pathogenesis. We also find that specific human leukocyte antigen (HLA) alleles form trans associations with the expression of AOAH and ARHGAP24 in monocytes but not in B cells. In summary, we show that mapping gene expression in defined primary cell populations identifies new cell type-specific trans-regulated networks and provides insights into the genetic basis of disease susceptibility.     

Mutations in MKKS cause obesity, retinal dystrophy and renal malformations associated with Bardet-Biedl syndrome

Bardet-Biedl syndrome (BBS) is an autosomal recessive disorder predominantly characterized by obesity, retinal dystrophy, polydactyly, learning difficulties, hypogenitalism and renal malformations, with secondary features that include diabetes mellitus, endocrinological dysfunction and behavioural abnormalities. Despite an initial expectation of genetic homogeneity due to relative clinical uniformity, five BBS loci have been reported, with evidence for additional loci in the human genome; however, no genes for BBS have yet been identified. We performed a genome screen with BBS families from Newfoundland that were excluded from BBS1-5 and identified linkage with D20S189. Fine-mapping reduced the critical interval to 1.9 cM between D20S851 and D20S189, encompassing a chaperonin-like gene. Mutations in this gene were recently reported to be associated with McKusick-Kaufman syndrome (MKKS; ref. 8). Given both the mapping position and clinical similarities of these two syndromes, we screened MKKS and identified mutations in five Newfoundland and two European-American BBS pedigrees. Most are frameshift alleles that are likely to result in a non-functional protein. Our data suggest that a complete loss of function of the MKKS product, and thus an inability to fold a range of target proteins, is responsible for the clinical manifestations of BBS.     

Adaptation shapes patterns of genome evolution on sexual and asexual chromosomes in Drosophila

What advantage might sexual recombination confer? Population genetics theory predicts that asexual genomes are less efficient at eliminating deleterious mutations and incorporating beneficial alleles. Here, I compare patterns of genome evolution in a 40-kb gene-rich region on homologous neo-sex chromosomes of Drosophila miranda. Genes on the non-recombining neo-Y show various signs of degeneration, including transposable-element insertions, frameshift mutations and a higher rate of amino-acid substitution. In contrast, loci on the recombining neo-X show intact open reading frames and generally low rates of amino-acid substitution. One exceptional gene on the neo-X shows evidence for adaptive protein evolution, affecting patterns of variability at neighboring regions along the chromosome. These findings illustrate the limits to natural selection in an asexual genome. Deleterious mutations, including repetitive DNA, accumulate on a non-recombining chromosome, whereas rapid protein evolution due to positive selection is confined to the recombining homolog.     

Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease

Integration of genome-wide expression profiling with linkage analysis is a new approach to identifying genes underlying complex traits. We applied this approach to the regulation of gene expression in the BXH/HXB panel of rat recombinant inbred strains, one of the largest available rodent recombinant inbred panels and a leading resource for genetic analysis of the highly prevalent metabolic syndrome. In two tissues important to the pathogenesis of the metabolic syndrome, we mapped cis- and trans-regulatory control elements for expression of thousands of genes across the genome. Many of the most highly linked expression quantitative trait loci are regulated in cis, are inherited essentially as monogenic traits and are good candidate genes for previously mapped physiological quantitative trait loci in the rat. By comparative mapping we generated a data set of 73 candidate genes for hypertension that merit testing in human populations. Mining of this publicly available data set is expected to lead to new insights into the genes and regulatory pathways underlying the extensive range of metabolic and cardiovascular disease phenotypes that segregate in these recombinant inbred strains.     

Detection of large-scale variation in the human genome

We identified 255 loci across the human genome that contain genomic imbalances among unrelated individuals. Twenty-four variants are present in > 10% of the individuals that we examined. Half of these regions overlap with genes, and many coincide with segmental duplications or gaps in the human genome assembly. This previously unappreciated heterogeneity may underlie certain human phenotypic variation and susceptibility to disease and argues for a more dynamic human genome structure.