Whole-Genome Sequencing Suggests Schizophrenia Risk Mechanisms in Humans with 22q11.2 Deletion Syndrome (original) (raw)

Abstract

Chromosome 22q11.2 microdeletions impart a high but incomplete risk for schizophrenia. Possible mechanisms include genome-wide effects of DGCR8 haploinsufficiency. In a proof-of-principle study to assess the power of this model, we used high-quality, whole-genome sequencing of nine individuals with 22q11.2 deletions and extreme phenotypes (schizophrenia, or no psychotic disorder at age >50 years). The schizophrenia group had a greater burden of rare, damaging variants impacting protein-coding neurofunctional genes, including genes involved in neuron projection (nominal P = 0.02, joint burden of three variant types). Variants in the intact 22q11.2 region were not major contributors. Restricting to genes affected by a DGCR8 mechanism tended to amplify between-group differences. Damaging variants in highly conserved long intergenic noncoding RNA genes also were enriched in the schizophrenia group (nominal P = 0.04). The findings support the 22q11.2 deletion model as a threshold-lowering first hit for schizophrenia risk. If applied to a larger and thus better-powered cohort, this appears to be a promising approach to identify genome-wide rare variants in coding and noncoding sequence that perturb gene networks relevant to idiopathic schizophrenia. Similarly designed studies exploiting genetic models may prove useful to help delineate the genetic architecture of other complex phenotypes.

Keywords: 22q11 deletion syndrome, next-generation sequencing, genetic architecture, copy number variation, microRNA, DGCR8, schizophrenia, noncoding RNA, lincRNA, FMR1, synapse, connectivity, postsynaptic density, polygenic risk score, ABLIM1, BSN, DIP2A, EXOC4, ITM2C, MYH9, MYH10, PCNT, PTPRG, SLITRK2, ZDHHC5

Schizophrenia is a complex neuropsychiatric disease with prominent genetic heterogeneity. The established molecular genetic risk factors of largest effect are rare copy number variations (CNVs), especially 22q11.2 deletions (Kirov et al. 2012; Costain et al. 2013; Stankiewicz and Lupski 2010; Lowther et al. 2015; Bassett et al. 2010; Hochstenbach et al. 2011; Costain and Bassett 2012; Rees et al. 2014). Whole-exome sequencing (WES) studies indicate that rare coding sequence variants also contribute to schizophrenia (Girard et al. 2011; Xu et al. 2012; Need et al. 2012; Gulsuner et al. 2013; Timms et al. 2013; Fromer et al. 2014; Purcell et al. 2014; McCarthy et al. 2014; Guipponi et al. 2014). Rare variants that disrupt mechanisms regulating expression of protein-coding genes are likely to be part of the genetic architecture of schizophrenia as well (Morrow 2015; Geaghan and Cairns 2014; Forstner et al. 2013; Beveridge and Cairns 2012; Moreau et al. 2011; Xu et al. 2010; Warnica et al. 2015). These variants usually alter noncoding RNA gene exons or splicing and transcription regulatory motifs that typically reside outside of protein-coding exons. For this reason, the majority of these variants are detectable only by the use of whole-genome sequencing (WGS).

Extensive research efforts have focused on understanding the contribution of common variation to schizophrenia risk. The most recent large-scale study successfully identified more than 100 genome-wide significant loci, although with very modest effect size and often obscure molecular mechanisms (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). A polygenic risk score, based on the additive contribution of many weakly associated variants (International Schizophrenia Consortium et al. 2009), has been used successfully to maximize the fraction of schizophrenia risk explained by common variation (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). Pathway-level methods also have been investigated to identify commonalities among many different contributing variants, rare or common (Kirov et al. 2012; Costain et al. 2013; Fromer et al. 2014; Purcell et al. 2014; Warnica et al. 2015; Pathway Analysis Subgroup of Psychiatric Genomics Consortium Network 2015).

Even in 22q11.2 deletion syndrome (22q11.2DS), where the recurrent 22q11.2 deletion imparts a 25% risk of developing schizophrenia (Fung et al. 2010; Schneider et al. 2014), there remain undiscovered determinants of expression. One proposed mechanism involves genome-wide microRNA (miRNA) dysregulation related to haploinsufficiency of the DGCR8 gene that lies within the 22q11.2 deletion region (Stark et al. 2008; Forstner et al. 2013; Schofield et al. 2011; Brzustowicz and Bassett 2012; Merico et al. 2014). In individuals with 22q11.2 deletions this haploinsufficiency could increase susceptibility to the effects of protein-coding mutations that otherwise may be tolerated, including those in genes that are involved in schizophrenia in the general population (Brzustowicz and Bassett 2012).

In this initial proof-of-principle study, we hypothesized that the 22q11.2 deletion would provide enhanced power to investigate biologically plausible mechanisms for schizophrenia. We used high-quality, WGS of nine individuals with 22q11.2 deletions and extreme phenotypes (schizophrenia, or no psychotic disorder at age >50 years), followed by a comprehensive annotation and prioritization of rare variants impacting coding and non-coding sequence (Yuen et al. 2015). To maximize statistical power, we investigated rare variant burden for gene-sets with higher a priori likelihood of contributing to schizophrenia risk. We additionally investigated common variant contribution using a polygenic risk score model.

We found evidence for rare variants outside the 22q11.2 region perturbing gene networks relevant to idiopathic schizophrenia, for a DGCR8/miRNA-related mechanism, for other noncoding sequence variants, and for a polygenic risk contribution, and predicted that maximal statistical power can be achieved with attainable sample sizes of this genetic model.

Methods and Materials

Subjects

From a cohort of Canadian adults with 22q11.2DS (Bassett et al. 2003, 2008; Brzustowicz and Bassett 2012; Cheung et al. 2014; Fung et al. 2010; Schneider et al. 2014; Vorstman et al. 2013; Swaby et al. 2011; Butcher et al. 2012, 2013, 2015), we selected nine unrelated individuals of European descent (Table 1), based on availability of high quality genomic DNA for WGS and phenotypic information consistent with the extreme phenotype design: six (SCZ1-SCZ6) met DSM-IV criteria for schizophrenia or schizoaffective disorder (Bassett et al. 2003) and three (NP1-NP3) had no psychotic disorder at age >50 years (Table 1). Deep phenotyping included direct clinical assessments at multiple time points and review of lifetime medical records, with the use of our established methods (Fung et al. 2010; Vorstman et al. 2013; Swaby et al. 2011; Butcher et al. 2012, 2013; Cheung et al. 2014; Bassett et al. 2003). The six subjects with schizophrenia had no other single major feature of 22q11.2DS in common (Table 1). All participants provided written informed consent, and the study was approved by local research ethics boards.

Table 1. Characteristics of nine adults of European ancestry with 22q11.2 deletions and whole-genome sequencing data.

WGS approach and methods

We submitted a high-quality genomic DNA sample from each subject to Complete Genomics for WGS (Drmanac et al. 2010; Carnevali et al. 2012). Mean genome coverage per sample was 98.95% (98.81–99.10%) at depth ≥5X and 97.65% (97.30–98.15%) at depth ≥10X, relative to the hg19 human genome reference sequence. In particular, 94.4% and 72.3% of the exome was covered with at least 20X and 40X sequence depth, respectively. Complete Genomics data for each nucleotide position, supplemented by in-house protocols, provided stringent quality filters. For this study, we used only high-quality variants (those with high confidence scores). Variants were then annotated with a custom pipeline based on the ANNOVAR (November 2014) software tool (Wang et al. 2010). We defined rare variants as those at <1% of the alternate allele frequency (minor allele frequency = 0.01) threshold in each of three standard [1000 Genomes (1000 Genomes Project Consortium et al. 2012), National Heart, Lung, and Blood Institute Exome Sequencing Project (Fu et al. 2013), Exome Aggregation Consortium (http://exac.broadinstitute.org/)], and two in-house, platform-matched databases. Details of all WGS-related laboratory and data interpretation/bioinformatics methods used are provided in the Supporting Information, Figure S1 and File S1.

Rare variant burden analyses for coding genes

We considered the possible impact of accumulated deleterious variants affecting protein-coding genes under a haploinsufficiency model, excluding variants in the intact chromosome 22q11.2 region and on the X chromosome, which were examined separately. These variants comprised three categories: loss of function (LoF) variants (stop-gain/nonsense, frameshift, and core splice site), damaging missense variants (predicted to be deleterious per five of seven standard tools), and splicing regulatory variants that (negatively) affect exon inclusion; the latter include intronic variants that are further away from core splice site (LoF) variants (Xiong et al. 2015). First, we tested “neurofunctional” gene-sets (i.e., affecting brain-related functions most likely to be important to schizophrenia expression), separating each variant category (LoF, missense, splicing regulatory). Gene-sets with nominally significant burden for at least one variant category (P < 0.10 for LoF and splicing regulatory, and P < 0.05 for missense variants) were then tested for the joint burden of the three variant categories with a multivariate, two-sample Hotelling’s T-Square test (Hotelling 1931). To investigate a DGCR8/miRNA mechanism, we used the same gene-sets but restricted to those genes predicted to be affected by DGCR8 haploinsufficiency (Stark et al. 2008). To estimate the burden effect size, we calculated the between-group ratio of the mean absolute variant count (Hu et al. 2009). For gene-set burden power calculations, we selected four representative gene-sets showing enrichment for one or more of the variant categories, and used Cohen’s d to express the effect size estimates.

Copy number variation

We evaluated CNVs and other structural variants (SVs) by using a previously established annotation and prioritization process (Yuen et al. 2015). All subjects were confirmed to have 22q11.2 deletions (Table 1). Of the remaining variants, only rare CNVs and SVs that overlapped at least one coding gene exon of a RefSeq gene with known neuronal function were considered in this study.

Rare variant burden analyses for noncoding RNA genes

We considered two main types of noncoding RNA variants: miRNA derived from mirBase v20 (Griffiths-Jones 2004) and long intergenic noncoding RNA (lincRNA) derived from the Broad catalog (Cabili et al. 2011). We tested the burden of high-quality, rare variants prioritized based on regional and nucleotide-level genomic conservation.

Common variant polygenic risk score

We obtained the list of 102,636 SNPs used by the Psychiatric Genomics Consortium to define a risk score for schizophrenia, together with the original nominal association p-values and odds ratios (International Schizophrenia Consortium et al. 2009; Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). These SNPs were mapped to hg19 coordinates and intersected with the WGS data for our cohort; in particular, WGS variants were matched to risk score SNPs by coordinates and alleles, whereas WGS reference intervals (i.e., identical to the human reference sequence) were matched by coordinate overlap. A total of 88,301 SNPs was successfully mapped to variants passing quality filters, or reference intervals, in all nine genomes in this study. Allele counts were computed as the number of alleles matching to the allele used for association analysis (possible values: 0, 1, 2). The SNP-wise risk score was then calculated as the product of this allele count and the log(odds-ratio) (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). Using nominal p-value thresholds ≤0.001 and ≤0.0001, as well as one more stringent (≤0.00001), and p-values >0.9 and >0.5 as negative controls (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014), the polygenic risk scores for each 22q11.2DS subject were then calculated as the sum of all respective SNP-wise risk scores (International Schizophrenia Consortium et al. 2009). Differences between schizophrenia and no-psychosis groups were tested using a one-sided _t_-test and a Wilcoxon test. We also calculated the percentage of correctly predicted schizophrenia and no-psychosis subjects at different risk score values, and reported the maximum value as a point-estimate of separation between the two groups.

Data availability

Supporting Information contains detailed descriptions of all supplemental files. Figure S1 contains selected gene-sets with a higher burden in subjects with schizophrenia. Figure S2 contains distribution boxplots of subjects' polygenic risk scores for the schizophrenia and nonpsychotic groups. Table S1 contains high quality, rare coding variants. Table S2 contains source and size of gene-sets used in the burden analyses. Table S3 contains details of burden analyses for each type of variants. Table S4 contains most recurrent splicing regulatory predictive features detected in this study. Table S5 contains details of power calculations. Table S6 contains details of burden analyses for lincRNA. Table S7 contains details of lincRNA with high quality, rare variants and miRNA with high quality rare variants.

Results

Details of subjects with 22q11.2DS are in Table 1. Subjects had an average of 13.8 and 94.3 high-quality, rare variants disrupting coding genes (LoF and missense categories, respectively), with similar findings for both the schizophrenia and non-psychotic groups (Table S3). There were few additional variants in the intact chromosome 22q11.2 region (Table 2).

Table 2. Selected brain function related gene-set results for rare single nucleotide variants.

Burden of rare variants impacting neurofunctional protein-coding genes

Table 2 shows all gene-sets with <2000 protein-coding genes and nominally significant (_P_ < 0.05, schizophrenia > nonpsychotic group) burden for rare deleterious variants. Only neurofunctional gene-sets met these criteria. On testing burden jointly for all three variant categories, only the Neuron projection [Gene Ontology (GO)] gene-set was significant (Hotelling’s T-Square P = 0.02). Table 2 shows the overlap between this and the other neurofunctional gene-sets for genes implicated in the schizophrenia group.

As predicted by a multiple within-person rare variant hypothesis for schizophrenia (Costain et al. 2013; Merico et al. 2014), there were several variants per subject involving these neurofunctional gene-sets. There were no significant between-group differences for larger gene-sets, or even all brain-expressed variants (Table S3). The findings support an approach focused on high-quality variants and gene-sets of neurofunctional relevance, even in this small sample. Table S1, Table S2, Table S3, and Table S4 show details for all high-quality, rare variants, gene-sets used, and burden analysis results.

We used the data available for the three variant types to perform power calculations for the Neuron projection (GO) gene-set burden test and three other gene-sets (Table S5). For N = 100 subjects per group, power for the GO gene-set was >0.99 for damaging missense variants and for LoF variants, and >0.94 for splicing regulatory variants (Cohen’s d effect sizes: 1.90, 0.88, 0.55, respectively; the effect size estimates are based on the nine genomes presented in this study). For the Post-synaptic density (Bayes et al. 2011) gene-set, power was >0.99 for LoF variants and for splicing regulatory variants. Other results showing power >0.99 are in Table S5.

Support for the DGCR8/miRNA hypothesis

Consistent with a miRNA hypothesis for schizophrenia, restricting to genes predicted to be affected by DGCR8 haploinsufficiency (Stark et al. 2008; Merico et al. 2014) tended to increase estimated effect sizes (Table 2 and Figure S1), despite the decrease in number of variants per subject. For missense variants, these gene-sets included Neuron projection (GO) and Synaptic pathways (Kyoto Encyclopedia of Genes and Genomes KEGG), with no overlap of the genes involved between these gene-sets. For LoF variants, the Post-synaptic density (Bayes et al. 2011) gene-set was implicated (Table 2). Restricting to _DGCR8_-related genes did not tend to increase effect size for splicing regulatory variants (Figure S1). Notably, applying the _DGCR8_-related gene filter revealed nominally significant burden in 22q11.2DS-schizophrenia for rare damaging missense variants using a gene-set from idiopathic schizophrenia WES studies (de novo nonsynonymous variants) (Girard et al. 2011; Xu et al. 2012; Gulsuner et al. 2013; Fromer et al. 2014; McCarthy et al. 2014; Guipponi et al. 2014). Several of the genes involved overlapped those in the Neuron projection (GO) gene-set (Table 2).

Rare CNV disrupting candidate genes for schizophrenia

Similar to our previous study focusing on CNV >10 kb in size (Bassett et al. 2008), we interrogated the genome outside of the 22q11.2 region for additional rare CNVs and SVs. In one individual with schizophrenia (SCZ4), we identified and confirmed via quantitative polymerase chain reaction a rare maternally inherited 84-kb deletion at 21q22.3. This CNV disrupts exons of the genes PCNT and DIP2A, the latter gene implicating DGCR8 and FMR1 interactome mechanisms (Stark et al. 2008; Darnell et al. 2011).

Rare variants disrupting noncoding RNA genes

There were multiple rare variants outside of protein-coding genes, on average involving 2.0 and 1.3 lincRNA genes per subject in the schizophrenia and nonpsychotic groups, respectively (Table S6). Restricting to highly conserved (top 10%) lincRNAs, the burden was greater in the schizophrenia group (mean 1.3 vs. 0; nominal P = 0.039) (Table S6). However, perhaps related to their small size, miRNA genes contained few rare variants, even after broadening the rarity definition to <5%, preventing statistical testing of burden (Table S7).

Schizophrenia polygenic risk score

Use of the selected schizophrenia-associated SNPs [at nominal p-value thresholds ≤0.001 and ≤0.0001 from the Psychiatric Consortium Study (International Schizophrenia Consortium et al. 2009; Schizophrenia Working Group Of The Psychiatric Genomics Consortium 2014)] resulted in greater polygenic risk scores in the 22q11.2DS schizophrenia than in the nonpsychotic group (for the two thresholds, respectively, based on 2866 and 1059 SNPs: means: 0.00798 vs. −2.482, 0.601 vs. −1.238; _t_-test p-values: P = 0.094, P = 0.064; Wilcoxon test p-values: P = 0.083, P = 0.190; maximum correctly predicted percentages: 83%, 75%) (Figure S2). These trends did not reach our definition of statistical significance, however. The Wilcoxon and _t_-test p-values were greater (0.136 and 0.274, respectively) using association threshold P ≤ 0.00001 (451 SNPs). As expected, almost no difference (Wilcoxon and _t_-test P = 0.32−0.80) was observed for negative control SNPs (18,675 and 1976 SNPs at association p-value thresholds P > 0.5 and P > 0.9, respectively).

Discussion

Historically, the psychiatric genetics field has not used a genetic model or a functionally and mechanistically driven approach as a means to evaluate germline genetic variation in schizophrenia. We demonstrate the potential success of exploiting the enhanced homogeneity and thus power of a genomic disorder (rare and highly penetrant CNV) to investigate expression of a major associated disease phenotype. This study of schizophrenia in 22q11.2DS had two primary goals: (i) to demonstrate an effective approach to analyzing and interpreting WGS data, based on previous success in autism (Yuen et al. 2015), and (ii) to identify and prioritize testable, biologically plausible hypotheses for further investigation in a larger sample. That there were findings reaching our definition of nominal statistical significance was unexpected in this small sample. The results demonstrate the power and generalizability of 22q11.2DS as a model for understanding the genetic architecture of idiopathic schizophrenia, and provide support for multiple rare variants within individuals and a miRNA-related mechanism. These are concepts with previous evidence (Girirajan et al. 2012; Warnica et al. 2015).

By definition, individuals with 22q11.2DS have a 22q11.2 deletion, thus identifying additional rare variants would support a multiple rare variant hypothesis for schizophrenia at the individual level. The findings of this study indicate that this is likely to involve not only exonic variants, as expected, but also variants in regulatory regions and noncoding RNA genes typically not detectable by WES technologies. In the subgroup of individuals with schizophrenia, there was evidence for enrichment of damaging variants in highly conserved lincRNA (nonprotein-coding) genes, and of certain splicing regulatory variants that affect protein-coding genes. Although functional characterization of lincRNAs is limited as yet, the strategy used here may help to identify lincRNAs that contribute to schizophrenia. lincRNAs are involved in epigenetic mechanisms including chromatin binding, and in splicing processes (Barry et al. 2014; Quek et al. 2015; Derrien et al. 2012; Moran et al. 2012). Interestingly, the gene-set most affected by splicing regulatory variants in this study implicates mRNA targets of FMR1, and thus post-transcriptional regulation of gene expression, including that involved in neuronal development and synaptic plasticity (Pinto et al. 2014; Suhl et al. 2014).

The burden analyses of the coding sequence variants further demonstrated the effectiveness of the approach used to analyze and interpret WGS data. In the subgroup of individuals with schizophrenia, using biologically informed filters revealed a greater burden of damaging variants affecting protein-coding genes involved in neuron projection (axonal and dendritic development), a gene-set previously implicated in schizophrenia using other approaches (Costain et al. 2013; Merico et al. 2014).

Restricting to genes affected by DGCR8 haploinsufficiency tended to increase effect sizes for neurofunctionally relevant gene-sets. The findings thus provide further support for a miRNA hypothesis for schizophrenia and the utility of 22q11.2DS as a model for this mechanism (Warnica et al. 2015; Merico et al. 2014; Morrow 2015; Geaghan and Cairns 2014; Moreau et al. 2011). The 22q11.2 deletion appears to act as a threshold-lowering first hit, likely in part related to haploinsufficiency of gene DGCR8 and its effects on miRNA buffering, to reveal effects of rare variants elsewhere in the genome (Stark et al. 2008; Forstner et al. 2013; Schofield et al. 2011; Brzustowicz and Bassett 2012; Merico et al. 2014). This included variants, present in each of the 22q11.2DS subjects with schizophrenia, in genes previously reported for idiopathic schizophrenia in WES studies (Girard et al. 2011; Xu et al. 2012; Gulsuner et al. 2013; Fromer et al. 2014; McCarthy et al. 2014; Guipponi et al. 2014).

Lastly, the polygenic risk score appears informative for the nine 22q11.2DS genomes, although probably because of the small sample size, the results do not achieve significance. Future studies with sufficient power to jointly model rare variant burden and common variant polygenic risk score would be useful, and could determine whether restricting the polygenic risk score SNPs to those implicating genes from neurofunctional gene-sets would amplify between-group differences.

Advantages and limitations

Although this initial study produced several nominally significant results, there was no correction for multiple comparisons. Part of our a priori design was that any findings would require replication with the use of larger samples. The estimates of effect size and power indicate that feasible sample sizes of individuals with 22q11.2DS will allow such replication, using a comparable design and approach. Our analytic strategy was designed to minimize both false-positive and false-negative results. In the absence of between-group differences in total burden of rare variants, individual false-positive results would be expected to affect both groups equally. All individuals would be expected to harbor multiple rare variants involved in neurofunctional gene-sets. Among genes in neurofunctional gene-sets, a specific subset may eventually be identified to make a greater contribution to the expression of schizophrenia in all, or in certain subforms, of the disorder. These could include genes where there are individual, rare damaging variants with large effect. Nonetheless, we expect a substantial level of polygenicity, as suggested by rare variant studies of schizophrenia and other neuropsychiatric disorders such as autism, as well as by the paucity of linkage findings for schizophrenia (Kirov et al. 2012; Costain et al. 2013; Girard et al. 2011; Xu et al. 2012; Need et al. 2012; Gulsuner et al. 2013; Timms et al. 2013; Fromer et al. 2014; Purcell et al. 2014; McCarthy et al. 2014; Guipponi et al. 2014; Pinto et al. 2014; Yuen et al. 2015).

As for the largest WES study in schizophrenia to date (Purcell et al. 2014), and our previous CNV studies (Pinto et al. 2014; Costain et al. 2013; Silversides et al. 2012), increased stringency of methods and approach, including quality, rarity, and deleteriousness of variants, generally strengthened the findings. Individual sequence variants were not validated by the use of a second method. Using a comparable WGS analytic pipeline, we found that greater than 90% of rare de novo SNVs were validated in a study of autism (Yuen et al. 2015); we expect this to be the minimum validation rate in this study. We would not restrict future studies to de novo variants, however, because most rare variants are inherited and may have enhanced impact in the context of a 22q11.2 deletion (Stark et al. 2008; Forstner et al. 2013; Schofield et al. 2011; Brzustowicz and Bassett 2012; Merico et al. 2014). For variants in nongenic regulatory regions, WGS is essential for detection with clear advantages over WES studies, including one involving two individuals with 22q11.2DS (Balan et al. 2014). Increasingly sophisticated methods offer promise for improved in silico evaluation of all variant types. Proof that a variant is causal, however, requires a laboratory-based functional analysis.

The size of this study limited the ability to explore interacting factors or to study other hypotheses of interest, including the role of individual common variants and nongenetic factors. Previous studies of SNPs on the intact chromosome 22q11.2 have shown conflicting or negative results, however reduced gene dosage of neurofunctional genes within the 22q11.2 region may contribute to schizophrenia risk (Philip and Bassett 2011; Arinami 2006; Prasad et al. 2008; Karayiorgou et al. 2010). Although we did not identify rare variants disrupting candidate genes in this region in individuals with schizophrenia, rare phenotypes may be attributable to unmasking of such variants (McDonald-McGinn et al. 2013). Our WGS approach could be generalizable to other phenotypes associated with 22q11.2DS, and help explain variable expression and incomplete penetrance in other genomic disorders. Also, although additional rare CNVs overlapping exons may play a minor role, consistent with previous results (Bassett et al. 2008; Williams et al. 2013), methods for studying small CNVs require refining. Nonetheless, the initial evidence of multiple rare variants within an individual, coupled with suggestive findings for polygenic common variant risk, provided by this study is consistent with a longstanding threshold model of schizophrenia.

Early studies of 22q11.2 deletions foreshadowed a more general role for rare CNV in understanding the global genetic architecture of schizophrenia in the population (Kirov et al. 2012; Costain et al. 2013; Stankiewicz and Lupski 2010; Lowther et al. 2015; Bassett et al. 2010; Hochstenbach et al. 2011; Costain and Bassett 2012; Zarrei et al. 2015; Rees et al. 2014). Here, results from this study indicate that, for the individual, to uncover the symphony of variants that increase the likelihood of the expression of schizophrenia, researchers can take advantage of the fact that a highly penetrant rare CNV like the 22q11.2 deletion represents an incomplete part of the genetic architecture. This study design may lead to the discovery of novel additional pathways from genotype to phenotype in schizophrenia. Findings from this study provide support for a tractable, mechanistically and functionally based approach for evaluating the myriad rare coding-sequence variants identified by WGS and their potential role along with common variant risk in individual genetic architecture, the importance of evaluating variants in noncoding sequence, and the enhanced power to identify relevant rare variants that may be afforded by a more genetically homogeneous sample. The main genetic sharing between unrelated individuals with schizophrenia appears to be at the pathway/mechanistic level, thus a research design with this focus promises robust findings. Studies that also combine common variant risk with a rare damaging variant gene-set burden model may be of particular interest.

Supplementary Material

Supporting Information

Acknowledgments

We thank the adults with 22q11.2DS and their families for their generous contributions to this and related research studies. The authors express gratitude to the students, research assistants, and staff affiliated with the Clinical Genetics Research Program and The Centre for Applied Genomics. Special thanks go to Thanuja Selvanayagam for assistance with laboratory experiments and to Monica Torsan for help with formatting. We thank Complete Genomics Inc. for providing the two Complete Genomics control data sets. We also thank the Exome Aggregation Consortium and the groups that provided exome variant data for comparison. A full list of contributing groups can be found at http://exac.broadinstitute.org/about. Finally, we thank Prof. Sarah Bergen for valuable suggestions for the calculation of the polygenic risk score. This work was supported by grants from Canadian Institutes of Health Research (CIHR) (MOP-97800 and MOP-89066), Genome Canada, and the University of Toronto McLaughlin Centre. S.W.S. holds the GlaxoSmithKline–CIHR Endowed Chair in Genome Sciences at The Hospital for Sick Children and the University of Toronto. A.S.B. holds the Canada Research Chair in Schizophrenia Genetics and Genomic Disorders, and the Dalglish Chair in 22q11.2 Deletion Syndrome.

Footnotes

Communicating editor: J. E. Richards

Literature Cited

1. Andrade D. M., Krings T., Chow E. W., Kiehl T. R., Bassett A. S., 2013. Hippocampal malrotation is associated with chromosome 22q11.2 microdeletion. Can. J. Neurol. Sci. 40: 652–656. [DOI] [PMC free article] [PubMed] [Google Scholar]
2. Arinami T., 2006. Analyses of the associations between the genes of 22q11 deletion syndrome and schizophrenia. J. Hum. Genet. 51: 1037–1045. [DOI] [PubMed] [Google Scholar]
3. Ascano M., Jr, Mukherjee N., Bandaru P., Miller J. B., Nusbaum J. D., et al. , 2012. FMRP targets distinct mRNA sequence elements to regulate protein expression. Nature 492: 382–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Balan S., Iwayama Y., Toyota T., Toyoshima M., Maekawa M., et al. , 2014. 22q11.2 deletion carriers and schizophrenia-associated novel variants. Br. J. Psychiatry 204: 398–399. [DOI] [PubMed] [Google Scholar]
5. Barry G., Briggs J. A., Vanichkina D. P., Poth E. M., Beveridge N. J., et al. , 2014. The long non-coding RNA Gomafu is acutely regulated in response to neuronal activation and involved in schizophrenia-associated alternative splicing. Mol. Psychiatry 19: 486–494. [DOI] [PubMed] [Google Scholar]
6. Bassett A. S., Chow E. W., AbdelMalik P., Gheorghiu M., Husted J., et al. , 2003. The schizophrenia phenotype in 22q11 deletion syndrome. Am. J. Psychiatry 160: 1580–1586. [DOI] [PMC free article] [PubMed] [Google Scholar]
7. Bassett A. S., Marshall C. R., Lionel A. C., Chow E. W., Scherer S. W., 2008. Copy number variations and risk for schizophrenia in 22q11.2 deletion syndrome. Hum. Mol. Genet. 17: 4045–4053. [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Bassett A. S., Scherer S. W., Brzustowicz L. M., 2010. Copy number variations in schizophrenia: critical review and new perspectives on concepts of genetics and disease. Am. J. Psychiatry 167: 899–914. [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Bayes A., van de Lagemaat L. N., Collins M. O., Croning M. D., Whittle I. R., et al. , 2011. Characterization of the proteome, diseases and evolution of the human postsynaptic density. Nat. Neurosci. 14: 19–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Beveridge N. J., Cairns M. J., 2012. MicroRNA dysregulation in schizophrenia. Neurobiol. Dis. 46: 263–271. [DOI] [PubMed] [Google Scholar]
11. Brzustowicz L. M., Bassett A. S., 2012. miRNA-mediated risk for schizophrenia in 22q11.2 deletion syndrome. Front. Genet. 3: 291. [DOI] [PMC free article] [PubMed] [Google Scholar]
12. Butcher N. J., Chow E. W., Costain G., Karas D., Ho A., et al. , 2012. Functional outcomes of adults with 22q11.2 deletion syndrome. Genet. Med. 14: 836–843. [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Butcher N., Kiehl T., Hazrati L., Chow E., Rogaeva E., et al. , 2013. Association between early-onset Parkinson disease and 22q11.2 deletion syndrome: identification of a novel genetic form of Parkinson disease and its clinical implications. JAMA Neurol. 70: 1359–1366. [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Butcher N., Marras C., Pondal M., Christopher L., Strafella A., et al. , 2014. Motor dysfunction in adults with hemizygous 22q11.2 deletions at high risk of early-onset Parkinson’s disease. Mov. Disord. 29(Suppl 1): S122. [Google Scholar]
15. Butcher N., Fung W., Fitzpatrick L., Guna A., Andrade D., et al. , 2015. Response to clozapine in a clinically identifiable subtype of schizophrenia. Br. J. Psychiatry 206: 484–491. [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Cabili M. N., Trapnell C., Goff L., Koziol M., Tazon-Vega B., et al. , 2011. Integrative annotation of human large intergenic noncoding RNAs reveals global properties and specific subclasses. Genes Dev. 25: 1915–1927. [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Carnevali P., Baccash J., Halpern A. L., Nazarenko I., Nilsen G. B., et al. , 2012. Computational techniques for human genome resequencing using mated gapped reads. J. Comput. Biol. 19: 279–292. [DOI] [PubMed] [Google Scholar]
18. Cheung E. N., George S. R., Costain G. A., Andrade D. M., Chow E. W., et al. , 2014. Prevalence of hypocalcemia and its associated features in 22q11.2 deletion syndrome. Clin. Endocrinol. (Oxf.) 81: 190–196. [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Costain G., Bassett A. S., 2012. Clinical applications of schizophrenia genetics: genetic diagnosis, risk, and counseling in the molecular era. Appl. Clin. Genet. 5: 1–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Costain G., Lionel A. C., Merico D., Forsythe P., Russell K., et al. , 2013. Pathogenic rare copy number variants in community-based schizophrenia suggest a potential role for clinical microarrays. Hum. Mol. Genet. 22: 4485–4501. [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Darnell J. C., Van Driesche S. J., Zhang C., Hung K. Y., Mele A., et al. , 2011. FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell 146: 247–261. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Derrien T., Johnson R., Bussotti G., Tanzer A., Djebali S., et al. , 2012. The GENCODE v7 catalog of human long noncoding RNAs: analysis of their gene structure, evolution, and expression. Genome Res. 22: 1775–1789. [DOI] [PMC free article] [PubMed] [Google Scholar]
23. Drmanac R., Sparks A. B., Callow M. J., Halpern A. L., Burns N. L., et al. , 2010. Human genome sequencing using unchained base reads on self-assembling DNA nanoarrays. Science 327: 78–81. [DOI] [PubMed] [Google Scholar]
24. Forstner A. J., Degenhardt F., Schratt G., Nothen M. M., 2013. MicroRNAs as the cause of schizophrenia in 22q11.2 deletion carriers, and possible implications for idiopathic disease: a mini-review. Front. Mol. Neurosci. 6: 47. [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Fromer M., Pocklington A. J., Kavanagh D. H., Williams H. J., Dwyer S., et al. , 2014. De novo mutations in schizophrenia implicate synaptic networks. Nature 506: 179–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Fu W., O’Connor T. D., Jun G., Kang H. M., Abecasis G., et al. , 2013. Analysis of 6,515 exomes reveals the recent origin of most human protein-coding variants. Nature 493: 216–220. [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Fung W. L., McEvilly R., Fong J., Silversides C., Chow E., et al. , 2010. Elevated prevalence of generalized anxiety disorder in adults with 22q11.2 deletion syndrome. Am. J. Psychiatry 167: 998. [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Fung W. L., Butcher N. J., Costain G., Andrade D. M., Boot E., et al. , 2015. Practical guidelines for managing adults with 22q11.2 deletion syndrome. Genet. Med. 17: 599–609. [DOI] [PMC free article] [PubMed] [Google Scholar]
29. Geaghan M., Cairns M. J., 2014. MicroRNA and posttranscriptional dysregulation in psychiatry. Biol. Psychiatry 78: 231–239. [DOI] [PubMed] [Google Scholar]
30. 1000 Genomes Project Consortium. Abecasis G. R., Auton A., Brooks L. D., DePristo M. A., Durbin R. M., et al. , 2012. An integrated map of genetic variation from 1,092 human genomes. Nature 491: 56–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Girard S. L., Gauthier J., Noreau A., Xiong L., Zhou S., et al. , 2011. Increased exonic de novo mutation rate in individuals with schizophrenia. Nat. Genet. 43: 860–863. [DOI] [PubMed] [Google Scholar]
32. Girirajan S., Rosenfeld J. A., Coe B. P., Parikh S., Friedman N., et al. , 2012. Phenotypic heterogeneity of genomic disorders and rare copy-number variants. N. Engl. J. Med. 367: 1321–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Griffiths-Jones S., 2004. The microRNA registry. Nucleic Acids Res. 32: D109–D111. [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Guipponi M., Santoni F. A., Setola V., Gehrig C., Rotharmel M., et al. , 2014. Exome sequencing in 53 sporadic cases of schizophrenia identifies 18 putative candidate genes. PLoS One 9: e112745. [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Gulsuner S., Walsh T., Watts A. C., Lee M. K., Thornton A. M., et al. , 2013. Spatial and temporal mapping of de novo mutations in schizophrenia to a fetal prefrontal cortical network. Cell 154: 518–529. [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Hochstenbach R., Buizer-Voskamp J. E., Vorstman J. A., Ophoff R. A., 2011. Genome arrays for the detection of copy number variations in idiopathic mental retardation, idiopathic generalized epilepsy and neuropsychiatric disorders: lessons for diagnostic workflow and research. Cytogenet. Genome Res. 135: 174–202. [DOI] [PubMed] [Google Scholar]
37. Hotelling H., 1931. The generalization of Student’s ratio. Ann. Math. Stat. 2: 360–378. [Google Scholar]
38. Hu P., Greenwood C. M., Beyene J., 2009. Using the ratio of means as the effect size measure in combining results of microarray experiments. BMC Syst. Biol. 3: 106. [DOI] [PMC free article] [PubMed] [Google Scholar]
39. International Schizophrenia Consortium. Purcell S. M., Wray N. R., Stone J. L., Visscher P. M., O’Donovan M. C., et al. , 2009. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature 460: 748–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Karayiorgou M., Simon T. J., Gogos J. A., 2010. 22q11.2 microdeletions: linking DNA structural variation to brain dysfunction and schizophrenia. Nat. Rev. Neurosci. 11: 402–416. [DOI] [PMC free article] [PubMed] [Google Scholar]
41. Kiehl T. R., Chow E. W., Mikulis D. J., George S. R., Bassett A. S., 2009. Neuropathologic features in adults with 22q11.2 deletion syndrome. Cereb. Cortex 19: 153–164. [DOI] [PubMed] [Google Scholar]
42. Kirov G., Pocklington A. J., Holmans P., Ivanov D., Ikeda M., et al. , 2012. De novo CNV analysis implicates specific abnormalities of postsynaptic signalling complexes in the pathogenesis of schizophrenia. Mol. Psychiatry 17: 142–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Lowther C., Costain G., Stavropoulos D. J., Melvin R., Silversides C. K., et al. , 2015. Delineating the 15q13.3 microdeletion phenotype: a case series and comprehensive review of the literature. Genet. Med. 17: 149–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
44. McCarthy S. E., Gillis J., Kramer M., Lihm J., Yoon S., et al. , 2014. De novo mutations in schizophrenia implicate chromatin remodeling and support a genetic overlap with autism and intellectual disability. Mol. Psychiatry 19: 652–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
45. McDonald-McGinn D. M., Fahiminiya S., Revil T., Nowakowska B. A., Suhl J., et al. , 2013. Hemizygous mutations in SNAP29 unmask autosomal recessive conditions and contribute to atypical findings in patients with 22q11.2DS. J. Med. Genet. 50: 80–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Merico D., Costain G., Butcher N. J., Warnica W., Ogura L., et al. , 2014. MicroRNA dysregulation, gene networks and risk for schizophrenia in 22q11.2 deletion syndrome. Front. Neurol. 5: 238. [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Moran V. A., Perera R. J., Khalil A. M., 2012. Emerging functional and mechanistic paradigms of mammalian long non-coding RNAs. Nucleic Acids Res. 40: 6391–6400. [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Moreau M. P., Bruse S. E., David-Rus R., Buyske S., Brzustowicz L. M., 2011. Altered microRNA expression profiles in postmortem brain samples from individuals with schizophrenia and bipolar disorder. Biol. Psychiatry 69: 188–193. [DOI] [PMC free article] [PubMed] [Google Scholar]
49. Morrow E. M., 2015. MicroRNAs in copy number variants in schizophrenia: misregulation of genome-wide gene expression programs. Biol. Psychiatry 77: 93–94. [DOI] [PubMed] [Google Scholar]
50. Need A. C., McEvoy J. P., Gennarelli M., Heinzen E. L., Ge D., et al. , 2012. Exome sequencing followed by large-scale genotyping suggests a limited role for moderately rare risk factors of strong effect in schizophrenia. Am. J. Hum. Genet. 91: 303–312. [DOI] [PMC free article] [PubMed] [Google Scholar]
51. Pathway Analysis Subgroup of Psychiatric Genomics Consortium Network , 2015. Psychiatric genome-wide association study analyses implicate neuronal, immune and histone pathways. Nat. Neurosci. 18: 199–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
52. Philip N., Bassett A., 2011. Cognitive, behavioural and psychiatric phenotype in 22q11.2 deletion syndrome. Behav. Genet. 41: 403–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
53. Pinto D., Delaby E., Merico D., Barbosa M., Merikangas A., et al. , 2014. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am. J. Hum. Genet. 94: 677–694. [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Piton A., Gauthier J., Hamdan F. F., Lafreniere R. G., Yang Y., et al. , 2011. Systematic resequencing of X-chromosome synaptic genes in autism spectrum disorder and schizophrenia. Mol. Psychiatry 16: 867–880. [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Prasad S. E., Howley S., Murphy K. C., 2008. Candidate genes and the behavioral phenotype in 22q11.2 deletion syndrome. Dev. Disabil. Res. Rev. 14: 26–34. [DOI] [PubMed] [Google Scholar]
56. Purcell S. M., Moran J. L., Fromer M., Ruderfer D., Solovieff N., et al. , 2014. A polygenic burden of rare disruptive mutations in schizophrenia. Nature 506: 185–190. [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Quek X. C., Thomson D. W., Maag J. L., Bartonicek N., Signal B., et al. , 2015. lncRNAdb v2.0: expanding the reference database for functional long noncoding RNAs. Nucleic Acids Res. 43: D168–D173. [DOI] [PMC free article] [PubMed] [Google Scholar]
58. Rees E., Walters J. T., Georgieva L., Isles A. R., Chambert K. D., et al. , 2014. Analysis of copy number variations at 15 schizophrenia-associated loci. Br. J. Psychiatry 204: 108–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Schizophrenia Working Group of the Psychiatric Genomics Consortium , 2014. Biological insights from 108 schizophrenia-associated genetic loci. Nature 511: 421–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
60. Schneider M., Debbane M., Bassett A. S., Chow E. W., Fung W. L., et al. , 2014. Psychiatric disorders from childhood to adulthood in 22q11.2 deletion syndrome: results from the International Consortium on Brain and Behavior in 22q11.2 Deletion Syndrome. Am. J. Psychiatry 171: 627–639. [DOI] [PMC free article] [PubMed] [Google Scholar]
61. Schofield C. M., Hsu R., Barker A. J., Gertz C. C., Blelloch R., et al. , 2011. Monoallelic deletion of the microRNA biogenesis gene Dgcr8 produces deficits in the development of excitatory synaptic transmission in the prefrontal cortex. Neural Dev. 6: 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
62. Silversides C. K., Lionel A. C., Costain G., Merico D., Migita O., et al. , 2012. Rare copy number variations in adults with tetralogy of Fallot implicate novel risk gene pathways. PLoS Genet. 8: e1002843. [DOI] [PMC free article] [PubMed] [Google Scholar]
63. Stankiewicz P., Lupski J. R., 2010. Structural variation in the human genome and its role in disease. Annu. Rev. Med. 61: 437–455. [DOI] [PubMed] [Google Scholar]
64. Stark K. L., Xu B., Bagchi A., Lai W. S., Liu H., et al. , 2008. Altered brain microRNA biogenesis contributes to phenotypic deficits in a 22q11-deletion mouse model. Nat. Genet. 40: 751–760. [DOI] [PubMed] [Google Scholar]
65. Suhl J. A., Chopra P., Anderson B. R., Bassell G. J., Warren S. T., 2014. Analysis of FMRP mRNA target datasets reveals highly associated mRNAs mediated by G-quadruplex structures formed via clustered WGGA sequences. Hum. Mol. Genet. 23: 5479–5491. [DOI] [PMC free article] [PubMed] [Google Scholar]
66. Swaby J. A., Silversides C. K., Bekeschus S. C., Piran S., Oechslin E. N., et al. , 2011. Complex congenital heart disease in unaffected relatives of adults with 22q11.2 deletion syndrome. Am. J. Cardiol. 107: 466–471. [DOI] [PMC free article] [PubMed] [Google Scholar]
67. Timms A. E., Dorschner M. O., Wechsler J., Choi K. Y., Kirkwood R., et al. , 2013. Support for the N-methyl-D-aspartate receptor hypofunction hypothesis of schizophrenia from exome sequencing in multiplex families. JAMA Psychiatry 70: 582–590. [DOI] [PubMed] [Google Scholar]
68. Vorstman J. A., Breetvelt E. J., Thode K. I., Chow E. W., Bassett A. S., 2013. Expression of autism spectrum and schizophrenia in patients with a 22q11.2 deletion. Schizophr. Res. 143: 55–59. [DOI] [PubMed] [Google Scholar]
69. Wang K., Li M., Hakonarson H., 2010. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 38: e164. [DOI] [PMC free article] [PubMed] [Google Scholar]
70. Warnica W., Merico D., Costain G., Alfred S. E., Wei J., et al. , 2015. Copy number variable microRNAs in schizophrenia and their neurodevelopmental gene targets. Biol. Psychiatry 77: 158–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
71. Williams H. J., Monks S., Murphy K. C., Kirov G., O’Donovan M. C., et al. , 2013. Schizophrenia two-hit hypothesis in velo-cardio facial syndrome. Am. J. Med. Genet. B. Neuropsychiatr. Genet. 162B: 177–182. [DOI] [PubMed] [Google Scholar]
72. Xiong H. Y., Alipanahi B., Lee L. J., Bretschneider H., Merico D., et al. , 2015. RNA splicing. The human splicing code reveals new insights into the genetic determinants of disease. Science 347: 1254806. [DOI] [PMC free article] [PubMed] [Google Scholar]
73. Xu B., Karayiorgou M., Gogos J. A., 2010. MicroRNAs in psychiatric and neurodevelopmental disorders. Brain Res. 1338: 78–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
74. Xu B., Ionita-Laza I., Roos J. L., Boone B., Woodrick S., et al. , 2012. De novo gene mutations highlight patterns of genetic and neural complexity in schizophrenia. Nat. Genet. 44: 1365–1369. [DOI] [PMC free article] [PubMed] [Google Scholar]
75. Yuen R. K., Thiruvahindrapuram B., Merico D., Walker S., Tammimies K., et al. , 2015. Whole-genome sequencing of quartet families with autism spectrum disorder. Nat. Med. 21: 185–191. [DOI] [PubMed] [Google Scholar]
76. Zarrei M., MacDonald J. R., Merico D., Scherer S. W., 2015. A copy number variation map of the human genome. Nat. Rev. Genet. 16: 172–183. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

Data Availability Statement

Supporting Information contains detailed descriptions of all supplemental files. Figure S1 contains selected gene-sets with a higher burden in subjects with schizophrenia. Figure S2 contains distribution boxplots of subjects' polygenic risk scores for the schizophrenia and nonpsychotic groups. Table S1 contains high quality, rare coding variants. Table S2 contains source and size of gene-sets used in the burden analyses. Table S3 contains details of burden analyses for each type of variants. Table S4 contains most recurrent splicing regulatory predictive features detected in this study. Table S5 contains details of power calculations. Table S6 contains details of burden analyses for lincRNA. Table S7 contains details of lincRNA with high quality, rare variants and miRNA with high quality rare variants.