Schizophrenia Research Forum - A Catalyst for Creative Thinking

WCPG 2012—Baby Steps for Sequencing in Schizophrenia

18 October 2012. More than 600 people gathered in Hamburg, Germany, 14-18 October, to attend the 20th World Congress of Psychiatric Genetics, organized by Markus Nöthen of the University of Bonn, Germany, and Marcella Rietschel of the University of Heidelberg in Mannheim, Germany. Buoyed by crisp fall weather and a boisterous reception with accordions and singing the night before, participants began Monday with two plenary talks that in some ways laid out the enormity of the task of psychiatric genetics: Karl Zilles of the University of Dusseldorf, Germany, displayed the complexities of human brain organization, and Mark Daly of the Broad Institute, Cambridge, Massachusetts, described the emerging genetic architecture of psychiatric disease.

Zilles showed off detailed regional specification in the brain, as revealed by neurotransmitter receptor location, and argued that taking these subcompartments into account will be crucial to understanding their function. He ended with a meditation on connectivity, with incredible views of axon structure afforded by polarized light microscopy, which can resolve single, myelinated axons.

Daly noted that recent developments in psychiatric genetics have been propelled by technical and collaborative advances alike. An example of this is an unprecedented number of researchers from around the world, sharing their schizophrenia and control samples to the Psychiatric Genomics Consortium (PGC). This gives more power to genomewide association studies (GWAS) to detect common variants contributing to the disorder (see SRF genetics series). Indeed, Daly announced 62 new genomewide significant hits found for schizophrenia in this “second wave” of PGC GWAS, which will be presented later in the meeting. “This is truly a profound moment in schizophrenia genetics,” he said. But he also tempered any unrealistic hopes for sequencing to make everything clear. Though sequencing is revealing many, many variants, they are rare, found in cases and controls alike, and scattered across many different genes. These challenges bring to mind the growing pains of early GWAS, and Daly counseled people to stay the course with sequencing, too, as a combination of common and rare variants will contribute to genetic risk for disease.

Lousy with variants
Against that cautionary backdrop on the challenges of interpreting sequencing results, researchers gathered in a session Monday afternoon to have a look at the latest sequencing results for schizophrenia. Michael O’Donovan of Cardiff University in the U.K. began by describing the hunt for de novo mutations—spontaneously occurring, non-inherited events—in the exomes of people with schizophrenia. Stressing the preliminary nature of the findings, he reported 485 de novo events in 586 parent-child trios from Bulgaria. The rates of mutation and the number of different types of mutation were similar to those found in previous de novo exome studies in schizophrenia (see SRF related news story) and autism (see SRF related news story), but unlike these previous studies, he and his colleagues have not found more de novo events in cases compared to controls. De novo events struck 12 genes in different people, and eight genes were hit by protein-altering non-silent changes. Though these “pileups” on single genes might look suspicious, the number was not greater than expected by chance. Taking sets of genes with related function as the unit of analysis, he found that schizophrenia cases had an enrichment for de novo events in postsynaptic density gene sets highlighted last year by de novo CNVs (see SRF related news story).

Continuing the search for de novo events in the exome (see SRF related news story), Guy Rouleau of the University of Montreal, Canada, presented findings from 14 more trios from France. This turned up 15 de novo events, four of which were protein-truncating nonsense mutations. When looking at the 30 genes hit by de novo events in his work and other published de novo studies (see SRF related news story), he found that they clustered in one pathway, which did not happen for 30 randomly chosen genes. Rouleau also noted an unnerving tendency for DNA derived from lymphoblastoid cell lines to harbor de novo events not found in DNA derived from blood cells, suggesting that DNA in the cell lines can mutate on its own.

Pinning a rare variant to disease might be easier if it occurred amid less genetic background noise, such as in genetically homogeneous isolated populations—and even better if the population happens to be enriched for disease-related variants. Aarno Palotie of the Wellcome Trust Sanger Institute in Cambridgeshire, U.K., may have both in a rural Finnish isolated population with increased prevalence of schizophrenia. Using exome sequencing, he reported that eight out of 22 families so far had novel loss-of-function mutations that segregate with disease. He suggested that the number of mutations coming out of this isolate might be enough to do association statistics to test their relation to schizophrenia.

Going the case-control route, Shaun Purcell of Mount Sinai School of Medicine in New York City gave a first-pass analysis of a massive haul of variants from exome sequencing of 5,023 Swedish samples, half with schizophrenia, half controls. Of the 60,000 variants identified, 2 percent were deemed loss of function. Cases did not carry an excess of loss-of-function variants over controls, but they did show a significant enrichment for loss-of-function variants in gene sets identified by previous GWAS, copy number variations (CNV), and de novo mutation studies, including the postsynaptic density one. Purcell also described other ideas for assessing a variant’s functional relevance, one of which involved paying attention to whether the variant hits the gene in a region that encodes a domain, a working part of a protein.

Mosaic of findings
On Tuesday, Richard McCombie of Cold Spring Harbor Laboratory, New York, presented his de novo events obtained from exome sequencing of 57 schizophrenia trios. Of 59 de novo mutations, some landed in genes already implicated in autism and intellectual disability. Noting that cases had an enrichment of de novo events in five chromatin-modifying genes (e.g., CHD8), he suggested that epigenetic control of chromatin might be disrupted in schizophrenia—something that could spur mutations in a subset of cells (“somatic mosaicism”) during development (Muotri et al., 2010).

Targeting the disrupted-in-schizophrenia 1 (DISC1) gene that emerged in a Scottish family beset by mental illness (see SRF related news story), David Porteous of the University of Edinburgh, U.K., gave an update on his quest to resequence the entire DISC1 gene and the flanking TRAX1 region containing the DISC1 promoter in 240 people with schizophrenia, 221 with bipolar disorder, 192 with recurrent major depression, and 889 healthy controls. He reported 145 exome variants, but these were not overrepresented in schizophrenia or bipolar cases compared to controls. One intronic variant, however, segregated in three families with recurrent major depressive disorder.

Though exome sequencing is producing results, it is still labor intensive and expensive. Much like the SNP chips of GWAS, an exome chip containing a wide variety of variants already discovered by sequencing could give a quicker and somewhat unbiased readout of this type of variation. In a different session on Tuesday, Ben Neale of Massachusetts General Hospital in Boston presented work in progress from an Illumina exome chip that he helped design, which contains about 250,000 variants derived from 12,000 individuals. Genotyping 5,196 people with schizophrenia and 6,500 controls, however, did not turn up an overrepresentation of any one of these variants in schizophrenia.—Michele Solis.

Comments on Related News

Related News: Autism Exome: Lessons for Schizophrenia?

Comment by:  Patrick Sullivan, SRF Advisor
Submitted 20 April 2012
Posted 23 April 2012
  I recommend the Primary Papers

Fascinating papers that likely presage work in the pipeline from multiple groups for schizophrenia. Truly groundbreaking work by some of the best groups in the business. Required reading for those interested in psychiatric genomics.

The identified loci provide important new windows into the neurobiology of ASD.

The results also pertain to the longstanding debate about the nature of ASD: does it result from many individually rare, Mendelian-like variants (potentially a different one in each person) and/or from the summation of the effects of many different common variants of subtle effects?

The multiple rare variant model now seems unlikely for ASD as, contrary to the expectations of some, ASD did not readily resolve into a handful of Mendelian-like diseases. (This comment is of course qualified by the limits of the technologies - which have, however, identified causal mutations for many monogenetic disorders.)

Readers might also want to read Ben Neale's comments on these papers at the Genomes Unzipped website.

View all comments by Patrick Sullivan

Related News: Exome Sequencing Hints at Prenatal Genes in Schizophrenia

Comment by:  Sven CichonMarcella RietschelMarkus M. Nöthen
Submitted 5 October 2012
Posted 5 October 2012

The new exome sequencing study by Xu et al. confirms previous results by the same research group (Xu et al., 2011) and by an independent group (Girard et al., 2011) that a significantly higher frequency of protein-altering de novo single nucleotide variants (SNVs) and in/dels is found in sporadic patients with schizophrenia. It is certainly reassuring that this observation has now been confirmed in an independent and considerably larger sample (134 patient-parent trios and 34 control-parent trios).

A closer look also reveals differences between this study and the study by Girard et al.: Xu et al. do not find a significantly higher overall de novo mutation rate per base per generation when comparing schizophrenia and control trios (1.73 x 10-08 vs. 1.28 x 10-08). In contrast, the Girard study found 2.59 x 10-08 de novo mutations in schizophrenia trios as opposed to the 1.1 x 10-08 events reported in the general population by the 1000 Genomes Project. The larger sample size in the new study by Xu et al., however, suggests that their estimation of the de novo mutation rates may be more precise now.

What eventually seems to count is the quality of the de novo mutations in the sporadic schizophrenia patients. The function of the genes hit by the non-synonymous/deleterious (as defined by in-silico scores) mutations is diverse and shows similarity with functions reported for common risk genes for schizophrenia identified by GWAS. Interestingly, there is an overrepresentation of genes that are predominantly expressed during embryogenesis, strongly highlighting a possible effect of neurodevelopmental disturbances in the etiology of schizophrenia (and nicely supporting what has already been concluded from GWAS).

It would probably be very interesting to estimate the penetrance of such de novo mutations to get a feeling for their individual impact on the development of the disease. In the absence of a reasonable number of individuals with the same mutation, however, this will be a difficult task.

Another aspect that is missing in the current paper, but is accessible to investigation, is the frequency/quality of de novo mutations in trios with a family history of schizophrenia and comparison to the figures seen in the sporadic trios. That might (or might not) support the authors’ conclusion that de novo events play a strong role in sporadic cases (and not in familial cases).


Xu B, Roos JL, Dexheimer P, Boone B, Plummer B, Levy S, Gogos JA, Karayiorgou M. Exome sequencing supports a de novo mutational paradigm for schizophrenia. Nat Genet . 2011 Sep ; 43(9):864-8. Abstract

Girard SL, Gauthier J, Noreau A, Xiong L, Zhou S, Jouan L, Dionne-Laporte A, Spiegelman D, Henrion E, Diallo O, Thibodeau P, Bachand I, Bao JY, Tong AH, Lin CH, Millet B, Jaafari N, Joober R, Dion PA, Lok S, Krebs MO, Rouleau GA. Increased exonic de novo mutation rate in individuals with schizophrenia. Nat Genet . 2011 Sep ; 43(9):860-3. Abstract

View all comments by Sven Cichon
View all comments by Marcella Rietschel
View all comments by Markus M. Nöthen

Related News: Exome Sequencing Hints at Prenatal Genes in Schizophrenia

Comment by:  Patrick Sullivan, SRF Advisor
Submitted 5 October 2012
Posted 5 October 2012

This paper by the productive group at Columbia increases our knowledge of the role of rare exon mutations in schizophrenia. The authors applied exome sequencing—a newish high-throughput sequencing technology—to trios consisting of both parents plus an offspring with schizophrenia. The authors focused on a subset of the genome (the “exome,” genetic regions believed to code for protein) on a subset of genetic variants (SNPs and insertion/deletion variants) of predicted functional significance, and on one type of inheritance (“de novo“ mutations, those absent in both parents and present in the offspring with schizophrenia).

The sample sizes are the largest yet reported for schizophrenia—231 affected trios and 34 controls. About 28 percent of these samples were reported in 2011 (Xu et al., 2011). A recent schizophrenia sequencing study (N = 166) from the Duke group was unrevealing (Need et al., 2012). The numbers in the Xu, 2012 paper are small compared to the three Nature trio studies for autism (see SRF related news story), an approximately threefold larger trio study for schizophrenia (in preparation), a case-control exome sequencing study for schizophrenia (total N ~5,000, in preparation), and a case-control exome chip study for schizophrenia (total N ~11,000, in preparation).

The authors reported:

more mutations with older fathers, as has been reported before (see SRF related news story). Note that advanced paternal age is an established risk factor for schizophrenia.

more de novo/predicted functional/exonic mutations in schizophrenia than in controls. However, the difference was slight, one-sided P = 0.03. One can quibble with the use of a one-tailed test (should never be used, in my opinion), but it is difficult to interpret this result unless paternal age is included as a covariate in this critical test.

an impressive set of bioinformatic and integrative analyses—see the paper for the large amount of work they did.

as might be predicted given the small sample size and the rarity of these sorts of mutations, there was no statistically significant pile-up of variants in specific genes. Hence, to my reading, the authors do not compellingly implicate any specific genes in the pathophysiology of schizophrenia. This conclusion is consistent with Need et al., 2012, and I note that the autism work implicated only a few genes (e.g., CHD8 and KATNAL2).

Note that the authors would disagree with the above, as they chose to focus on a set of genes that they thought stood out (reporting an aggregate P of 0.002), and the last third of the paper focuses on these genes. However, the human genetics community now insists on two critical points for implicating specific genes in associations with a disorder. The first is statistical significance, and the critical P value for an exome sequencing study is on the order of 1E-6. The second is replication. In my view, neither of these standards are achieved. However, their observations are intriguing, and may well eventually move us forward.

The key observation in this paper is the increased rate of de novo variation in schizophrenia cases. Is the increased rate indeed part of an etiological process? In other words, older fathers have an increased chance of exonic mutations, and these, in turn, increase risk for schizophrenia? Or are these merely hitch-hikers of no particularly biological import?

A major issue with exome studies is that there are so many predicted functional variants in apparently normal people. We all carry on the order of 100 exonic variants of predicted functional consequences with on the order of 20 genes that are probable knockouts. If part of the risk for schizophrenia indeed resides in the exome, very large studies will be required to identify such loci confidently. Moreover, published work on autism and unpublished work for type 2 diabetes, coronary artery disease, and schizophrenia suggest that this will require very large sample sizes, on the order of 100 times more than reported here. And, it is possible that the exome is not all that important for schizophrenia.


Xu B, Roos JL, Dexheimer P, Boone B, Plummer B, Levy S, Gogos JA, Karayiorgou M. Exome sequencing supports a de novo mutational paradigm for schizophrenia. Nat Genet . 2011 Sep ; 43(9):864-8. Abstract

Need AC, McEvoy JP, Gennarelli M, Heinzen EL, Ge D, Maia JM, Shianna KV, He M, Cirulli ET, Gumbs CE, Zhao Q, Campbell CR, Hong L, Rosenquist P, Putkonen A, Hallikainen T, Repo-Tiihonen E, Tiihonen J, Levy DL, Meltzer HY, Goldstein DB. 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 . 2012 Aug 10 ; 91(2):303-12. Abstract

View all comments by Patrick Sullivan