Search

The success of next-generation sequencing depends on the accuracy of variant calls. Few objective protocols exist for QC following variant calling from whole genome sequencing (WGS) data. After applying QC filtering based on Genome Analysis Tool Kit (GATK) best practices, we used genotype discordance of eight samples that were sequenced twice each to evaluate the proportion of potentially inaccurate variant calls. We designed a QC pipeline involving hard filters to improve replicate genotype concordance, which indicates improved accuracy of genotype calls. Our pipeline analyzes the efficacy of each filtering step. We initially applied this strategy to well-characterized variants from the ClinVar database, and subsequently to the full WGS dataset. The genome-wide biallelic pipeline removed 82.11% of discordant and 14.89% of concordant genotypes, and improved the concordance rate from 98.53% to 99.69%. The variant-level read depth filter most improved the genome-wide biallelic concordance rate. We also adapted this pipeline for triallelic sites, given the increasing proportion of multiallelic sites as sample sizes increase. For triallelic sites containing only SNVs, the concordance rate improved from 97.68% to 99.80%. Our QC pipeline removes many potentially false positive calls that pass in GATK, and may inform future WGS studies prior to variant effect analysis.

Androgenetic alopecia (AGA), or male-pattern baldness, is the most common form of hair loss. Its pathogenesis is androgen dependent, and genetic predisposition is the major requirement for the phenotype. We demonstrate that genetic variability in the androgen receptor gene (AR) is the cardinal prerequisite for the development of early-onset AGA, with an etiological fraction of 0.46. The investigation of a large number of genetic variants covering the AR locus suggests that a polyglycine-encoding GGN repeat in exon 1 is a plausible candidate for conferring the functional effect. The X-chromosomal location of AR stresses the importance of the maternal line in the inheritance of AGA.

In order to identify single-nucleotide polymorphisms (SNPs) and analyze their characteristics in a set of 111 genes, we resequenced exons and flanking regions in an average of 170 chromosomes from individuals of European origin. Genetic variability was decreased in noncoding regions highly conserved between human and rodents, indicating functional relevance of these regions. Furthermore, diversity of coding nonsynonymous SNPs was found lower in regions encoding a known protein sequence motif. SNPs predicted to be of functional significance were more common amongst rare variants. Despite the significant recent growth of SNP numbers in public SNP databases, only a small fraction of these rare variants is represented. This may be relevant in the investigation of the genetic causes of severe side effects, for which rare variants are plausible candidates. Estimation of htSNPs reduces the genotyping effort required in capturing common haplotypes, for certain genes, however, this accounts for only a small fraction of haplotype diversity.

We compared coding region variants of 53 cognitively healthy centenarians and 45 patients with Alzheimer’s disease (AD), all of Ashkenazi Jewish (AJ) ancestry. Despite the small sample size, the known AD risk variant APOE4 reached genome-wide significance, indicating the advantage of utilizing ‘super-controls’. We restricted our subsequent analysis to rare variants observed at most once in the 1000 Genomes database and having a minor allele frequency below 2% in our AJ sample. We compared the burden of predicted protein altering variants between cases and controls as normalized by the level of rare synonymous variants. We observed an increased burden among AD subjects for predicted loss-of-function (LoFs) variants defined as stop-gain, frame shift, initiation codon (INIT) and splice site mutations (n = 930, OR = 1.3, P = 1.5×E−5). There was no enrichment across all rare protein altering variants defined as missense plus LoFs, in frame indels and stop-loss variants (n = 13 014, OR = 0.97, P = 0.47). Among LoFs, the strongest burden was observed for INIT (OR = 2.16, P = 0.0097) and premature stop variants predicted to cause non-sense-mediated decay in the majority of transcripts (NMD) (OR = 1.98, P = 0.02). Notably, this increased burden of NMD, INIT and splice variants was more pronounced in a set of 1397 innate immune genes (OR = 4.55, P = 0.0043). Further comparison to additional exomes indicates that the difference in LoF burden originated both from the AD and centenarian sample. In summary, we observed an overall increased burden of rare LoFs in AD subjects as compared to centenarians, and this enrichment is more pronounced for innate immune genes.

The neurotransmitter serotonin [5-hydroxytryptamine (5-HT)] controls a broad range of biological functions that are disturbed in affective disorder. In the brain, 5-HT production is controlled by tryptophan hydroxylase 2 (TPH2). In order to assess the possible contribution of TPH2 genetic variability to the aetiology of bipolar affective disorder (BPAD), we systematically investigated common and rare genetic variation in the TPH2 gene through a sequential sequencing and SNP-based genotyping approach. Our study sample comprised two cohorts of BPAD from Germany and Russia, totalling 883 patients and 1300 controls. SNPs located in a haplotype block covering the 5' region of the gene as well as a rare, non-synonymous SNP, resulting in a Pro206Ser substitution, showed significant association with bipolar disorder. The odds ratio for the minor allele in the pooled sample was 1.5 (95% CI 1.2-1.9) for rs11178997 (in the 5'-associated haplotype block) and 4.8 (95% CI 1.6-14.8) for rs17110563 encoding the Pro206Ser substitution. Examination of the functional effects of TPH2 Pro206Ser provided evidence for a reduced thermal stability and solubility of the mutated enzyme, suggesting reduced 5-HT production in the brain as a pathophysiological mechanism in BPAD.

A recent study has suggested that the g.961C>G (p.Ser278Arg) variant of the autoimmune regulator (AIRE) gene contributes to susceptibility to alopecia areata (AA). We attempted to replicate this finding using a case-control sample of Belgian-German origin (273 patients and 283 controls). Despite adequate power, our study results do not support a significant association of the risk allele in our AA patient sample. This remained the case when we stratified our sample according to severity and family history of disease. Our study results do not support the hypothesis that the g.961C>G (p.Ser278Arg) polymorphism of the AIRE gene is associated with an increased risk for AA.

In order to explore the role of noncoding variants in the genetics of schizophrenia, we sequenced 27 kb of noncoding DNA from the gene loci RAC-alpha serine/threonine-protein kinase (AKT1), brain-derived neurotrophic factor (BDNF), dopamine receptor-3 (DRD3), dystrobrevin binding protein-1 (DTNBP1), neuregulin-1 (NRG1) and regulator of G-protein signaling-4 (RGS4) in 37 schizophrenia patients and 25 healthy controls. To compare the allele frequency spectrum between the two samples, we separately computed Tajima's D-value for each sample. The results showed a smaller Tajima's D-value in the case sample, pointing to an excess of rare variants as compared to the control sample. When randomly permuting the affection status of sequenced individuals, we observed a stronger decrease of Tajima's D in 2400 out of 100 000 permutations, corresponding to a P-value of 0.024 in a one-sided test. Thus, rare variants are significantly enriched in the schizophrenia sample, indicating the existence of disease-related sequence alterations. When categorizing the sequenced fragments according to their level of human–rodent conservation or according to their gene locus, we observed a wide range of diversity parameter estimates. Rare variants were enriched in conserved regions as compared to nonconserved regions in both samples. Nevertheless, rare variants remained more common among patients, suggesting an increased number of variants under purifying selection in this sample. Finally, we performed a heuristic search for the subset of gene loci, which jointly produces the strongest difference between controls and cases. This showed a more prominent role of variants from the loci AKT1, BDNF and RGS4. Taken together, our approach provides promising strategy to investigate the genetics of schizophrenia and related phenotypes.

The detailed investigation of variation in functionally important regions of the human genome is expected to promote understanding of genetically complex diseases. We resequenced 65 candidate genes for CNS disorders in an average of 85 European individuals. The minor allele frequency (MAF), an indicator of weak purifying selection, was lowest in radical amino acid alterations, whereas similar MAF was observed for synonymous variants and conservative amino acid alterations. In noncoding sequences, variants located in CpG islands tended to have a lower MAF than those outside CpG islands. The transition/transversion ratio was increased among both synonymous and conservative variants compared with noncoding variants. Conversely, the transition/transversion ratio was lowest among radical amino acid alterations. Furthermore, among nonsynonymous variants, transversions displayed lower MAF than did transitions. This suggests that transversions are associated with functionally important amino acid alterations. By comparing our data with public SNP databases, we found that variants with lower allele frequency are underrepresented in these databases. Therefore, radical variants obtain distinctively lower database coverage. However, those variants appear to be under weak purifying selection and thus could play a role in the etiology of genetically complex diseases.

To measure the strength of natural selection that acts upon single nucleotide variants (SNVs) in a set of human genes, we calculate the ratio between nonsynonymous SNVs (nsSNVs) per nonsynonymous site and synonymous SNVs (sSNVs) per synonymous site. We transform this ratio with a respective factor f that corrects for the bias of synonymous sites towards transitions in the genetic code and different mutation rates for transitions and transversions. This method approximates the relative density of nsSNVs (rdnsv) in comparison with the neutral expectation as inferred from the density of sSNVs. Using SNVs from a diploid genome and 200 exomes, we apply our method to immune system genes (ISGs), nervous system genes (NSGs), randomly sampled genes (RSGs), and gene ontology annotated genes. The estimate of rdnsv in an individual exome is around 20% for NSGs and 30–40% for ISGs and RSGs. This smaller rdnsv of NSGs indicates overall stronger purifying selection. To quantify the relative shift of nsSNVs towards rare variants, we next fit a linear regression model to the estimates of rdnsv over different SNV allele frequency bins. The obtained regression models show a negative slope for NSGs, ISGs and RSGs, supporting an influence of purifying selection on the frequency spectrum of segregating nsSNVs. The y-intercept of the model predicts rdnsv for an allele frequency close to 0. This parameter can be interpreted as the proportion of nonsynonymous sites where mutations are tolerated to segregate with an allele frequency notably greater than 0 in the population, given the performed normalization of the observed nsSNV to sSNV ratio. A smaller y-intercept is displayed by NSGs, indicating more nonsynonymous sites under strong negative selection. This predicts more monogenically inherited or de-novo mutation diseases that affect the nervous system.