The contribution of de novo coding mutations to autism spectrum disorder.

Whole exome sequencing has proven to be a powerful tool for understanding the genetic architecture of human disease. Here we apply it to more than 2,500 simplex families, each having a child with an autistic spectrum disorder. By comparing affected to unaffected siblings, we show that 13% of de novo missense mutations and 43% of de novo likely gene-disrupting (LGD) mutations contribute to 12% and 9% of diagnoses, respectively. Including copy number variants, coding de novo mutations contribute to about 30% of all simplex and 45% of female diagnoses. Almost all LGD mutations occur opposite wild-type alleles. LGD targets in affected females significantly overlap the targets in males of lower intelligence quotient (IQ), but neither overlaps significantly with targets in males of higher IQ. We estimate that LGD mutation in about 400 genes can contribute to the joint class of affected females and males of lower IQ, with an overlapping and similar number of genes vulnerable to contributory missense mutation. LGD targets in the joint class overlap with published targets for intellectual disability and schizophrenia, and are enriched for chromatin modifiers, FMRP-associated genes and embryonically expressed genes. Most of the significance for the latter comes from affected females.

A tumour suppressor network relying on the polyamine-hypusine axis.

Tumour suppressor genes encode a broad class of molecules whose mutational attenuation contributes to malignant progression. In the canonical situation, the tumour suppressor is completely inactivated through a two-hit process involving a point mutation in one allele and chromosomal deletion of the other. Here, to identify tumour suppressor genes in lymphoma, we screen a short hairpin RNA library targeting genes deleted in human lymphomas. We functionally identify those genes whose suppression promotes tumorigenesis in a mouse lymphoma model. Of the nine tumour suppressors we identified, eight correspond to genes occurring in three physically linked 'clusters', suggesting that the common occurrence of large chromosomal deletions in human tumours reflects selective pressure to attenuate multiple genes. Among the new tumour suppressors are adenosylmethionine decarboxylase 1 (AMD1) and eukaryotic translation initiation factor 5A (eIF5A), two genes associated with hypusine, a unique amino acid produced as a product of polyamine metabolism through a highly conserved pathway. Through a secondary screen surveying the impact of all polyamine enzymes on tumorigenesis, we establish the polyamine-hypusine axis as a new tumour suppressor network regulating apoptosis. Unexpectedly, heterozygous deletions encompassing AMD1 and eIF5A often occur together in human lymphomas and co-suppression of both genes promotes lymphomagenesis in mice. Thus, some tumour suppressor functions can be disabled through a two-step process targeting different genes acting in the same pathway.

Patterns and rates of exonic de novo mutations in autism spectrum disorders.

Autism spectrum disorders (ASD) are believed to have genetic and environmental origins, yet in only a modest fraction of individuals can specific causes be identified. To identify further genetic risk factors, here we assess the role of de novo mutations in ASD by sequencing the exomes of ASD cases and their parents (n = 175 trios). Fewer than half of the cases (46.3%) carry a missense or nonsense de novo variant, and the overall rate of mutation is only modestly higher than the expected rate. In contrast, the proteins encoded by genes that harboured de novo missense or nonsense mutations showed a higher degree of connectivity among themselves and to previous ASD genes as indexed by protein-protein interaction screens. The small increase in the rate of de novo events, when taken together with the protein interaction results, are consistent with an important but limited role for de novo point mutations in ASD, similar to that documented for de novo copy number variants. Genetic models incorporating these data indicate that most of the observed de novo events are unconnected to ASD; those that do confer risk are distributed across many genes and are incompletely penetrant (that is, not necessarily sufficient for disease). Our results support polygenic models in which spontaneous coding mutations in any of a large number of genes increases risk by 5- to 20-fold. Despite the challenge posed by such models, results from de novo events and a large parallel case-control study provide strong evidence in favour of CHD8 and KATNAL2 as genuine autism risk factors.

Duplications of the neuropeptide receptor gene VIPR2 confer significant risk for schizophrenia.

Rare copy number variants (CNVs) have a prominent role in the aetiology of schizophrenia and other neuropsychiatric disorders. Substantial risk for schizophrenia is conferred by large (>500-kilobase) CNVs at several loci, including microdeletions at 1q21.1 (ref. 2), 3q29 (ref. 3), 15q13.3 (ref. 2) and 22q11.2 (ref. 4) and microduplication at 16p11.2 (ref. 5). However, these CNVs collectively account for a small fraction (2-4%) of cases, and the relevant genes and neurobiological mechanisms are not well understood. Here we performed a large two-stage genome-wide scan of rare CNVs and report the significant association of copy number gains at chromosome 7q36.3 with schizophrenia. Microduplications with variable breakpoints occurred within a 362-kilobase region and were detected in 29 of 8,290 (0.35%) patients versus 2 of 7,431 (0.03%) controls in the combined sample. All duplications overlapped or were located within 89 kilobases upstream of the vasoactive intestinal peptide receptor gene VIPR2. VIPR2 transcription and cyclic-AMP signalling were significantly increased in cultured lymphocytes from patients with microduplications of 7q36.3. These findings implicate altered vasoactive intestinal peptide signalling in the pathogenesis of schizophrenia and indicate the VPAC2 receptor as a potential target for the development of new antipsychotic drugs.

Mapping copy number variation by population-scale genome sequencing.

Genomic structural variants (SVs) are abundant in humans, differing from other forms of variation in extent, origin and functional impact. Despite progress in SV characterization, the nucleotide resolution architecture of most SVs remains unknown. We constructed a map of unbalanced SVs (that is, copy number variants) based on whole genome DNA sequencing data from 185 human genomes, integrating evidence from complementary SV discovery approaches with extensive experimental validations. Our map encompassed 22,025 deletions and 6,000 additional SVs, including insertions and tandem duplications. Most SVs (53%) were mapped to nucleotide resolution, which facilitated analysing their origin and functional impact. We examined numerous whole and partial gene deletions with a genotyping approach and observed a depletion of gene disruptions amongst high frequency deletions. Furthermore, we observed differences in the size spectra of SVs originating from distinct formation mechanisms, and constructed a map of SV hotspots formed by common mechanisms. Our analytical framework and SV map serves as a resource for sequencing-based association studies.

Finding disease variants in Mendelian disorders by using sequence data: methods and applications.

Many sequencing studies are now underway to identify the genetic causes for both Mendelian and complex traits. Via exome-sequencing, genes harboring variants implicated in several Mendelian traits have already been identified. The underlying methodology in these studies is a multistep algorithm based on filtering variants identified in a small number of affected individuals and depends on whether they are novel (not yet seen in public resources such as dbSNP), shared among affected individuals, and other external functional information on the variants. Although intuitive, these filter-based methods are nonoptimal and do not provide any measure of statistical uncertainty. We describe here a formal statistical approach that has several distinct advantages: (1) it provides fast computation of approximate p values for individual genes, (2) it adjusts for the background variation in each gene, (3) it allows for incorporation of functional or linkage-based information, and (4) it accommodates designs based on both affected relative pairs and unrelated affected individuals. We show via simulations that the proposed approach can be used in conjunction with the existing filter-based methods to achieve a substantially better ranking of a gene relevant for disease when compared to currently used filter-based approaches, this is especially so in the presence of disease locus heterogeneity. We revisit recent studies on three Mendelian diseases and show that the proposed approach results in the implicated gene being ranked first in all studies, and approximate p values of 10(-6) for the Miller Syndrome gene, 1.0 × 10(-4) for the Freeman-Sheldon Syndrome gene, and 3.5 × 10(-5) for the Kabuki Syndrome gene.

Whole exome sequencing reveals minimal differences between cell line and whole blood derived DNA.

Two common sources of DNA for whole exome sequencing (WES) are whole blood (WB) and immortalized lymphoblastoid cell line (LCL). However, it is possible that LCLs have a substantially higher rate of mutation than WB, causing concern for their use in sequencing studies. We compared results from paired WB and LCL DNA samples for 16 subjects, using LCLs of low passage number (<5). Using a standard analysis pipeline we detected a large number of discordant genotype calls (approximately 50 per subject) that we segregated into categories of "confidence" based on read-level quality metrics. From these categories and validation by Sanger sequencing, we estimate that the vast majority of the candidate differences were false positives and that our categories were effective in predicting valid sequence differences, including LCLs with putative mosaicism for the non-reference allele (3-4 per exome). These results validate the use of DNA from LCLs of low passage number for exome sequencing.

AnnTools: a comprehensive and versatile annotation toolkit for genomic variants.

UNLABELLED: AnnTools is a versatile bioinformatics application designed for comprehensive annotation of a full spectrum of human genome variation: novel and known single-nucleotide substitutions (SNP/SNV), short insertions/deletions (INDEL) and structural variants/copy number variation (SV/CNV). The variants are interpreted by interrogating data compiled from 15 constantly updated sources. In addition to detailed functional characterization of the coding variants, AnnTools searches for overlaps with regulatory elements, disease/trait associated loci, known segmental duplications and artifact prone regions, thereby offering an integrated and comprehensive analysis of genomic data. The tool conveniently accepts user-provided tracks for custom annotation and offers flexibility in input data formats. The output is generated in the universal Variant Call Format. High annotation speed makes AnnTools suitable for high-throughput sequencing facilities, while a low-memory footprint and modest CPU requirements allow it to operate on a personal computer. The application is freely available for public use; the package includes installation scripts and a set of helper tools. AVAILABILITY: http://anntools.sourceforge.net/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Detecting common copy number variants in high-throughput sequencing data by using JointSLM algorithm.

The discovery of genomic structural variants (SVs), such as copy number variants (CNVs), is essential to understand genetic variation of human populations and complex diseases. Over recent years, the advent of new high-throughput sequencing (HTS) platforms has opened many opportunities for SVs discovery, and a very promising approach consists in measuring the depth of coverage (DOC) of reads aligned to the human reference genome. At present, few computational methods have been developed for the analysis of DOC data and all of these methods allow to analyse only one sample at time. For these reasons, we developed a novel algorithm (JointSLM) that allows to detect common CNVs among individuals by analysing DOC data from multiple samples simultaneously. We test JointSLM performance on synthetic and real data and we show its unprecedented resolution that enables the detection of recurrent CNV regions as small as 500 bp in size. When we apply JointSLM to analyse chromosome one of eight genomes with different ancestry, we identify 3000 regions with recurrent CNVs of different frequency and size: hierarchical clustering on these regions segregates the eight individuals in two groups that reflect their ancestry, demonstrating the potential utility of JointSLM for population genetics studies.

Sharing parental genomes by siblings concordant or discordant for autism.

Studying thousands of families, we find siblings concordant for autism share more of their parental genomes than expected by chance, and discordant siblings share less, consistent with a role of transmission in autism incidence. The excess sharing of the father is highly significant (p value of 0.0014), with less significance for the mother (p value of 0.31). To compare parental sharing, we adjust for differences in meiotic recombination to obtain a p value of 0.15 that they are shared equally. These observations are contrary to certain models in which the mother carries a greater load than the father. Nevertheless, we present models in which greater sharing of the father is observed even though the mother carries a greater load. More generally, our observations of sharing establish quantitative constraints that any complete genetic model of autism must satisfy, and our methods may be applicable to other complex disorders.

Exome sequencing identifies GCDH (glutaryl-CoA dehydrogenase) mutations as a cause of a progressive form of early-onset generalized dystonia.

Dystonias are a clinically and genetically heterogeneous group of movement disorders characterized by involuntary, sustained muscular contractions affecting one or more sites of the body, and abnormal postures. In this study, we describe an autosomal recessive family that presents with a progressive and early-onset form of generalized dystonia. The nuclear family consists of two healthy parents and two affected daughters. To elucidate the genetic causes underlying disease, whole-exome sequencing analysis was performed in one affected sibling, followed by validation, biochemical analyses and MRI brain imaging. A homozygous, disease-segregating mutation (p.Val400Met) was identified in the glutaryl-CoA dehydrogenase (GCDH) gene at chromosome 19p13. The mutation, in an amino acid that is highly conserved among species, was absent in large number of neurologically normal individuals. Biochemical analyses demonstrated increased 3-hydroxy glutaric acid present in urine samples from both patients. MRI imaging revealed a T2 and flair hyperintense signal in lenticular nuclei with bilateral and symmetrical distribution. We conclude that both GCDH activity and GCDH mutation analysis should be considered in the differential diagnosis of progressive forms of early-onset generalized dystonia and that mitochondrial fatty acid metabolism is one important pathway in the development of dystonia. As lysine restriction and L: -carnitine supplementation are important treatments for GCDH deficiency, identification of this deficiency in patients with progressive forms of early-onset generalized dystonia has potential treatment implications.

Sensitive and accurate detection of copy number variants using read depth of coverage.

Methods for the direct detection of copy number variation (CNV) genome-wide have become effective instruments for identifying genetic risk factors for disease. The application of next-generation sequencing platforms to genetic studies promises to improve sensitivity to detect CNVs as well as inversions, indels, and SNPs. New computational approaches are needed to systematically detect these variants from genome sequence data. Existing sequence-based approaches for CNV detection are primarily based on paired-end read mapping (PEM) as reported previously by Tuzun et al. and Korbel et al. Due to limitations of the PEM approach, some classes of CNVs are difficult to ascertain, including large insertions and variants located within complex genomic regions. To overcome these limitations, we developed a method for CNV detection using read depth of coverage. Event-wise testing (EWT) is a method based on significance testing. In contrast to standard segmentation algorithms that typically operate by performing likelihood evaluation for every point in the genome, EWT works on intervals of data points, rapidly searching for specific classes of events. Overall false-positive rate is controlled by testing the significance of each possible event and adjusting for multiple testing. Deletions and duplications detected in an individual genome by EWT are examined across multiple genomes to identify polymorphism between individuals. We estimated error rates using simulations based on real data, and we applied EWT to the analysis of chromosome 1 from paired-end shotgun sequence data (30x) on five individuals. Our results suggest that analysis of read depth is an effective approach for the detection of CNVs, and it captures structural variants that are refractory to established PEM-based methods.

Rates of contributory de novo mutation in high and low-risk autism families.

Autism arises in high and low-risk families. De novo mutation contributes to autism incidence in low-risk families as there is a higher incidence in the affected of the simplex families than in their unaffected siblings. But the extent of contribution in low-risk families cannot be determined solely from simplex families as they are a mixture of low and high-risk. The rate of de novo mutation in nearly pure populations of high-risk families, the multiplex families, has not previously been rigorously determined. Moreover, rates of de novo mutation have been underestimated from studies based on low resolution microarrays and whole exome sequencing. Here we report on findings from whole genome sequence (WGS) of both simplex families from the Simons Simplex Collection (SSC) and multiplex families from the Autism Genetic Resource Exchange (AGRE). After removing the multiplex samples with excessive cell-line genetic drift, we find that the contribution of de novo mutation in multiplex is significantly smaller than the contribution in simplex. We use WGS to provide high resolution CNV profiles and to analyze more than coding regions, and revise upward the rate in simplex autism due to an excess of de novo events targeting introns. Based on this study, we now estimate that de novo events contribute to 52-67% of cases of autism arising from low risk families, and 30-39% of cases of all autism.

A Bayesian approach for applying Haseman-Elston methods.

The main goal of this paper is to couple the Haseman-Elston method with a simple yet effective Bayesian factor-screening approach. This approach selects markers by considering a set of multigenic models that include epistasis effects. The markers are ranked based on their marginal posterior probability. A significant improvement over our previously proposed Bayesian variable selection methodology is a simple Metropolis-Hasting algorithm that requires minimum tuning on the prior settings. The algorithm, however, is also flexible enough for us to easily incorporate our hypotheses and avoid computational pitfalls. We apply our approach to the microsatellite data of Collaborative Studies on Genetics of Alcoholism using the coded values for the ALDX1 variable as our response.

Characterization of SLITRK1 variation in obsessive-compulsive disorder.

Obsessive compulsive disorder (OCD) is a syndrome characterized by recurrent and intrusive thoughts and ritualistic behaviors or mental acts that a person feels compelled to perform. Twin studies, family studies, and segregation analyses provide compelling evidence that OCD has a strong genetic component. The SLITRK1 gene encodes a developmentally regulated stimulator of neurite outgrowth and previous studies have implicated rare variants in this gene in disorders in the OC spectrum, specifically Tourette syndrome (TS) and trichotillomania (TTM). The objective of the current study was to evaluate rare genetic variation in SLITRK1 in risk for OCD and to functionally characterize associated coding variants. We sequenced SLITRK1 coding exons in 381 individuals with OCD as well as in 356 control samples and identified three novel variants in seven individuals. We found that the combined mutation load in OCD relative to controls was significant (p = 0.036). We identified a missense N400I change in an individual with OCD, which was not found in more than 1000 control samples (P<0.05). In addition, we showed the the N400I variant failed to enhance neurite outgrowth in primary neuronal cultures, in contrast to wildtype SLITRK1, which enhanced neurite outgrowth in this assay. These important functional differences in the N400I variant, as compared to the wildtype SLITRK1 sequence, may contribute to OCD and OC spectrum symptoms. A synonymous L63L change identified in an individual with OCD and an additional missense change, T418S, was found in four individuals with OCD and in one individual without an OCD spectrum disorder. Examination of additional samples will help assess the role of rare SLITRK1 variation in OCD and in related psychiatric illness.

Identification of genes and variants associated with quantitative traits using Bayesian factor screening.

We propose a factor-screening method based on a Bayesian model selection framework and apply it to Genetic Analysis Workshop 17 simulated data with unrelated individuals to identify genes and SNP variants associated with the quantitative trait Q1. A Metropolis-Hasting algorithm is implemented to generate a posterior distribution in a restricted model space and thus the marginal posterior distribution of each variant. Our framework provides flexibility to make inferences on either individual variants or genes. We obtained results for 10 simulated data sets. Our methods are able to identify FTP1 and KDR, two genes that are associated with Q1 in a majority of replicates.

Principal components ancestry adjustment for Genetic Analysis Workshop 17 data.

Statistical tests on rare variant data may well have type I error rates that differ from their nominal levels. Here, we use the Genetic Analysis Workshop 17 data to estimate type I error rates and powers of three models for identifying rare variants associated with a phenotype: (1) by using the number of minor alleles, age, and smoking status as predictor variables; (2) by using the number of minor alleles, age, smoking status, and the identity of the population of the subject as predictor variables; and (3) by using the number of minor alleles, age, smoking status, and ancestry adjustment using 10 principal component scores. We studied both quantitative phenotype and a dichotomized phenotype. The model with principal component adjustment has type I error rates that are closer to the nominal level of significance of 0.05 for single-nucleotide polymorphisms (SNPs) in noncausal genes for the selected phenotype than the model directly adjusting for population. The principal component adjustment model type I error rates are also closer to the nominal level of 0.05 for noncausal SNPs located in causal genes for the phenotype. The power for causal SNPs with the principal component adjustment model is comparable to the power of the other methods. The power using the underlying quantitative phenotype is greater than the power using the dichotomized phenotype.

Inferring haplotypes of copy number variations from high-throughput data with uncertainty.

Accurate information on haplotypes and diplotypes (haplotype pairs) is required for population-genetic analyses; however, microarrays do not provide data on a haplotype or diplotype at a copy number variation (CNV) locus; they only provide data on the total number of copies over a diplotype or an unphased sequence genotype (e.g., AAB, unlike AB of single nucleotide polymorphism). Moreover, such copy numbers or genotypes are often incorrectly determined when microarray signal intensities derived from different copy numbers or genotypes are not clearly separated due to noise. Here we report an algorithm to infer CNV haplotypes and individuals' diplotypes at multiple loci from noisy microarray data, utilizing the probability that a signal intensity may be derived from different underlying copy numbers or genotypes. Performing simulation studies based on known diplotypes and an error model obtained from real microarray data, we demonstrate that this probabilistic approach succeeds in accurate inference (error rate: 1-2%) from noisy data, whereas previous deterministic approaches failed (error rate: 12-18%). Applying this algorithm to real microarray data, we estimated haplotype frequencies and diplotypes in 1486 CNV regions for 100 individuals. Our algorithm will facilitate accurate population-genetic analyses and powerful disease association studies of CNVs.

Rare de novo and transmitted copy-number variation in autistic spectrum disorders.

To explore the genetic contribution to autistic spectrum disorders (ASDs), we have studied genomic copy-number variation in a large cohort of families with a single affected child and at least one unaffected sibling. We confirm a major contribution from de novo deletions and duplications but also find evidence of a role for inherited "ultrarare" duplications. Our results show that, relative to males, females have greater resistance to autism from genetic causes, which raises the question of the fate of female carriers. By analysis of the proportion and number of recurrent loci, we set a lower bound for distinct target loci at several hundred. We find many new candidate regions, adding substantially to the list of potential gene targets, and confirm several loci previously observed. The functions of the genes in the regions of de novo variation point to a great diversity of genetic causes but also suggest functional convergence.

High frequencies of de novo CNVs in bipolar disorder and schizophrenia.

While it is known that rare copy-number variants (CNVs) contribute to risk for some neuropsychiatric disorders, the role of CNVs in bipolar disorder is unclear. Here, we reasoned that a contribution of CNVs to mood disorders might be most evident for de novo mutations. We performed a genome-wide analysis of de novo CNVs in a cohort of 788 trios. Diagnoses of offspring included bipolar disorder (n = 185), schizophrenia (n = 177), and healthy controls (n = 426). Frequencies of de novo CNVs were significantly higher in bipolar disorder as compared with controls (OR = 4.8 [1.4,16.0], p = 0.009). De novo CNVs were particularly enriched among cases with an age at onset younger than 18 (OR = 6.3 [1.7,22.6], p = 0.006). We also confirmed a significant enrichment of de novo CNVs in schizophrenia (OR = 5.0 [1.5,16.8], p = 0.007). Our results suggest that rare spontaneous mutations are an important contributor to risk for bipolar disorder and other major neuropsychiatric diseases.