Data availability De novo variants discovered from the new trios are published in Supplementary Table S12. The data that support the findings of this study are available from the corresponding author upon request. Code availability A description of the R functions used for statistical analysis can be found in the Life Sciences Reporting Summary. Author contributions MCOD, MJO, JTRW, PH and ER conceived and designed the research. ER analysed the data. JH, JM and NC performed and managed the sequencing experiments. JH and MD performed the Sanger sequencing validation experiment. VEP, AJP, LH, SEL, AFP and ALR contributed to the interpretation of the results.

Schizophrenia is a highly polygenic disorder with important contributions coming from both common and rare risk alleles, the latter including CNVs and rare coding variants (RCVs), sometimes occurring as de novo variants (DNVs). We performed DNV analysis in whole exome-sequencing data obtained from a new sample of 613 schizophrenia trios, and combined this with published data for a total of 3,444 trios. Loss-of-function (LoF) DNVs were significantly enriched among 3,488 LoF intolerant genes in our new trio data (rate ratio (RR) (95% CI) = 2.23 (1.31, 3.79); p = 2.2 × 10−3), supporting previous findings. In the full dataset, genes associated with neurodevelopmental disorders (NDD; n=160) were significantly enriched for LoF DNVs (RR (95% CI) = 3.32 (2.0, 5.21); p = 7.4 × 10−6). Within this set of NDD genes, SLC6A1, encoding a gamma-aminobutyric acid transporter, was associated with missense-damaging DNVs (p = 5.2 × 10−5). Using data from a subset of 1,122 trios for which we had genome-wide common variant data, schizophrenia polygenic risk was significantly over-transmitted to probands (p = 2.6 × 10−60), as was bipolar disorder common variant polygenic risk (p = 5.7 × 10−17). We defined carriers of candidate schizophrenia-related DNVs as those with LoF or deletion DNVs in LoF intolerant or NDD genes. These individuals had significantly less over-transmission of common risk alleles than non-carriers (p = 3.5 × 10−4), providing robust support for the hypothesis that this set of DNVs is enriched for those related to schizophrenia.

Recent advances in psychiatric genetics have led to the discovery of dozens of genomic loci associated with schizophrenia. However, a gap exists between the detection of genetic associations and understanding the underlying molecular mechanisms. This review describes the basic approaches used in the so-called post-GWAS studies to generate biological interpretation of the existing population genetic data, including both molecular (creation and analysis of knockout animals, exploration of the transcriptional effects of common variants in human brain cells) and computational (fine-mapping of causal variability, gene set enrichment analysis, partitioned heritability analysis) methods. The results of the crucial studies, in which these approaches were used to uncover the molecular and neurobiological basis of the disease, are also reported.

BackgroundMethylation of DNA is associated with a variety of biological processes. With whole-genome studies of DNA methylation, it became possible to determine a set of genomic sites where DNA methylation is associated with a specific phenotype. A method is needed that allows detailed follow-up studies of the sites, including taking into account genetic information. Bisulfite PCR is a natural choice for this kind of task, but multiplexing is one of the most important problems impeding its implementation. To address this task, we took advantage of a recently published method based on Pacbio sequencing of long bisulfite PCR products (single-molecule real-time bisulfite sequencing, SMRT-BS) and tested the validity of the improved methodology with a smoking phenotype.ResultsHerein, we describe the “panhandle” modification of the method, which permits a more robust PCR with multiple targets. We applied this technique to determine smoking by DNA methylation in 71 healthy people and 83 schizophrenia patients (n = 50 smokers and n = 104 non-smokers, Russians of the Moscow region). We used five targets known to be influenced by smoking (regions of genes AHRR, ALPPL2, IER3, GNG12, and GFI1). We discovered significant allele-specific methylation effects in the AHRR and IER3 regions and assessed how this information could be exploited to improve the prediction of smoking based on the collected DNA methylation data. We found no significant difference in the methylation profiles of selected targets in relation to schizophrenia suggesting that smoking affects methylation at the studied genomic sites in healthy people and schizophrenia patients in a similar way.ConclusionsWe determined that SMRT-BS with “panhandle” modification performs well in the described setting. Additional information regarding methylation and allele-specific effects could improve the predictive accuracy of DNA methylation-based models, which could be valuable for both basic research and clinical applications.Electronic supplementary materialThe online version of this article (10.1186/s13148-018-0565-1) contains supplementary material, which is available to authorized users.

Large-scale epigenomic projects have mapped hundreds of thousands of potential regulatory sites in the human genome, but only a small proportion of these elements are proximal to transcription start sites. It is believed that the majority of these sequences are remote promoter-activating genomic sites scattered within several hundreds of kilobases from their cognate promoters and referred to as enhancers. It is still unclear what principles, aside from relative closeness in the linear genome, determine which promoter(s) is controlled by a given enhancer; however, this understanding is of great fundamental and clinical relevance. In recent years, C-methods (chromosome conformation capture-based methods) have become a powerful tool for the identification of enhancer–promoter spatial contacts that, in most cases, reflect their functional link. Here, we describe a new hybridisation-based promoter Capture-C protocol that makes use of biotinylated dsDNA probes generated by PCR from a custom pool of long oligonucleotides. The described protocol allows high-resolution promoter interactome description, providing a flexible and cost-effective alternative to the existing promoter Capture-C modifications. Based on the obtained data, we propose several tips on probe design that could potentially improve the results of future experiments.

Interrogating DNA methylation within schizophrenia risk loci holds promise to identify mechanisms by which genes influence the disease. Based on the hypothesis that allele specific methylation (ASM) of a single CpG, or perhaps CpH, might mediate or mark the effects of genetic variants on disease risk and phenotypes, we explored haplotype specific methylation levels of individual cytosines within a genomic region harbouring the BAG5, APOPT1 and KLC1 genes in peripheral blood of schizophrenia patients and healthy controls. Three DNA fragments located in promoter, intronic and intergenic areas were studied by single-molecule real-time bisulfite sequencing enabling the analysis of long reads of DNA with base-pair resolution and the determination of haplotypes directly from sequencing data. Among 1,012 cytosines studied, we did not find any site where methylation correlated with the disease or cognitive deficits after correction for multiple testing. At the same time, we determined the methylation profile associated with the schizophrenia risk haplotype within the KLC1 fourth intron and confirmed ASM for cytosines located in the vicinity of rs67899457. These genetically associated DNA methylation variations may be related to the pathophysiological mechanism differentiating the risk and non-risk haplotypes and merit further investigation.

No abstract

Scientifically interesting as well as practically important phenotypes often belong to the realm of complex traits. To the extent that these traits are hereditary, they are usually ‘highly polygenic’. The study of such traits presents a challenge for researchers, as the complex genetic architecture of such traits makes it nearly impossible to utilise many of the usual methods of reverse genetics, which often focus on specific genes. In recent years, thousands of genome-wide association studies (GWAS) were undertaken to explore the relationships between complex traits and a large number of genetic factors, most of which are characterised by tiny effects. In this review, we aim to familiarise ‘wet biologists’ with approaches for the interpretation of GWAS results, to clarify some issues that may seem counterintuitive and to assess the possibility of using GWAS results in experiments on various complex traits.