Email Icon Facebook icon Twitter Icon GooglePlus Icon Contact

User Top Menu

WCPG 2015—Schizophrenia Insights Through Combining Common and Rare Variants

25 Oct 2015

October 26, 2015. Though separate analyses point to rare and common genetic contributions to psychiatric disorders, a session on Monday, October 19, explored how the two combined. Similar to a recent paper (see SRF related news report), speakers reported some support for a liability threshold model in which rare alleles with large effects and common alleles with small effects add up to reach a disease burden.

Presenting data from the Psychiatric Genomic Consortium's (PGC) schizophrenia cases, Sarah Bergen of the Karolinska Institute, Sweden, reported that the polygenic risk score—a measure of the aggregated burden of schizophrenia risk alleles a person has—was lower in people who carried copy number variants (CNVs) than in non-carriers. A CNV-specific analysis found that the polygenic risk score varied by the potency of the CNV: For example, those with 15q11 deletions had higher polygenic risk scores than those carrying 22q11.2 deletions, which have the highest effects on risk. This suggests that cases with a powerful risk factor require less risk from the rest of the genome to reach the threshold for disease. Bergen said that she was currently exploring whether common variants varied with the specific features of schizophrenia.

In an afternoon session on schizophrenia genetics, Pippa Thomson of the University of Edinburgh, U.K., gave an update on the famous Scottish DISC1 family first identified 15 years ago (see SRF related news report). The family carries a translocation that swaps pieces of chromosome 11 with chromosome 1—a genetic glitch that interrupts the DISC1 gene, which has since been associated with brain development and synapse function. Thomson gave an update on the family, which now includes data from 105 people spanning six generations. The relative proportions of mental illness in the family (schizophrenia, major depressive disorder (MDD), and bipolar disorder) have stayed the same, and carrying the translocation was still highly associated with these disorders. Some family members without the translocation, however, were also affected by MDD and anxiety. Whole-genome sequencing of 49 individuals, combined with linkage analysis, detected signals at chromosomes 1 and 11, consistent with the translocation, as well as at loci on chromosomes 3 and 5. Thomson suggested that these new loci may explain the variable presentation of mental illness in the family. She also noted that the polygenic risk score for schizophrenia alleles was higher in schizophrenia and bipolar disorder, but not MDD, compared to unaffected family members.

George Kirov of Cardiff University, U.K., presented a follow-up analysis of 15q11 duplication CNVs that have been associated with schizophrenia when they are inherited from the mother (see SRF related news report). Expanding this analysis to 28,000 people with schizophrenia, 100,000 with developmental disorders including autism, and 150,000 controls, Kirov reported that the association between 15q11 duplications inherited from the mother and schizophrenia remained firm. These maternal duplications were found more frequently in schizophrenia than were paternal duplications; however, the rate of maternal duplications in developmental disorders was nearly twice that found in schizophrenia. Kirov said that schizophrenia cases with these maternal duplications were characterized by high catatonia, disorganized behavior, and a low intelligence quotient (IQ).

Though the latest genomewide association study (GWAS) made great headway in identifying risk loci for schizophrenia (see SRF related news report), researchers remain unsure about how the identified alleles confer risk. Some of the GWAS-identified SNPs may be dials that powerfully regulate gene expression, according to data presented by Menachem Fromer of the Icahn School of Medicine at Mount Sinai, New York City. Fromer reported on data from the CommonMind Consortium's effort to systematically map the molecular landscape of hundreds of postmortem brains, focusing on RNA sequencing results from 467 samples of the dorsolateral prefrontal cortex (dlPFC). This identified genomic loci that vary with gene expression ("eQTLs"), and some of these overlapped with some of the 108 GWAS-identified SNPs. He highlighted five genes regulated by GWAS-SNPs, including FURIN, TSNARE1, and CNTN4.

Although CNVs escalate risk for schizophrenia, they often disrupt multiple genes, making it hard to pinpoint which one is most important to risk. Douglas Ruderfer of the Icahn School of Medicine at Mount Sinai, New York City, presented a method for inferring the most damaging genes, based on simply observing how often a gene was hit by a CNV. If hit less often than expected, then that gene would be deemed "intolerant" to CNV mutation, and more likely be deleterious. Using exome sequencing data from the Exome Aggregation Consortium (ExAC) database of 60,000 samples, Ruderfer calculated a gene's intolerance. CNVs in schizophrenia cases landed in regions measured to be more intolerant than those hit in controls.

The session finished up with what may have been an audio-visual first for the meeting: a pre-recorded slide show and talk given by Evangelos Vassos of King's College London, who was recovering in London from injuries sustained while rescuing a kite from a tree! Studying 712 first-episode psychosis cases, Vassos reported that polygenic risk scores for schizophrenia risk alleles were significantly higher in these early cases than in controls among Europeans, but not for Africans. Among Europeans, polygenic risk scores were higher in those with psychosis due to schizophrenia than in those with psychosis for other reasons. Vassos, and his colleague Marta Di Forti, who answered questions on his behalf, suggested that the polygenic risk score might eventually help with diagnosis, since only 50 percent of people get a diagnosis at their first episode of psychosis. But at this point, it did not have the needed specificity and selectivity to accurately make this discrimination.—Michele Solis.