In a recent paper, Peykov et al.1 tried to find evidence for the role of SHANK2 in schizophrenia (SCZ) susceptibility by sequencing and experimental analysis. The authors claim that ‘results strongly suggest a causative role of rare SHANK2 variants in SCZ’. Unfortunately, this conclusion is not well supported by their data.
The first putative evidence is an excess of carriers of rare missense variants in patients versus controls (P=0.039). Nevertheless, they considered rare variants as those at minor allele frequency (MAF) in controls <1%. The use of MAF in controls for definition of rare variants is a well-known source of bias.2, 3 By this way, there is an upper limit for frequency in controls but not in patients, which may lead to an increase of type I errors. In fact, this is the case of the Peykov et al.’s data.1 Significance of the results depends on the exclusion of p.R569H, present in 16 of 659 controls (MAF=1.2%) and 4 of 481 cases (MAF=0.42%). Using the unbiased criteria of MAF<1% in the combined sample, there are 37/481 carriers in SCZ patients and 42/659 carriers in controls (χ2-test P=0.386).
An additional argument used by Peykov et al.1 as evidence of a role for SHANK2 in SCZ susceptibility is the existence of a variant, p.A1731S, in four cases and its absence in 5338 controls of European ancestry (P=4.6 × 10−5, two-sided Fisher’s exact test). These controls are the 659 sequenced controls from Germany (374) and from the Ontario Population Genomics Project control collection (285),4 4300 European American from the NHLBI GO Exome Sequencing Project and 379 European samples from The 1000 Genomes Project from diverse populations such as British in England and Scotland, Iberian in Spain, Toscani in Italia, Finnish and the CEU (Centre d’Etude du Polymorphisme Humain, UT) samples.5 However, rare variants analysis is especially susceptible to recent population history. For instance, The 1000 Genomes Project analyzed a total of 14 populations, including the five of European ancestry cited above. As much as 53% of rare variants at 0.5% were observed in a single population.5 Therefore, careful matching of samples by geographical origin is fundamental, and a test comparing the frequency of a rare variant in patients at a specific population versus the frequency found in a set of different European populations is not a strong evidence for involvement of a rare variant in a disease. In fact, comparison of carrier frequencies between all the samples reported in Peykov et al. irrespective of case–control status and the samples from the public databases used by them detected two significant associations, the claimed high-risk variant p.A1731S (P=1.5 × 10−3, two-sided Fisher’s exact test) and p.Y967C (P=5.5 × 10−5, two-sided Fisher’s exact test), strongly suggesting population stratification.
A final result used to claim SHANK2 involvement in SCZ is based on experimental evidence. The authors performed a series of elegant experiments in cultured primary hippocampal neurons from rat and fibroblast cell lines to test for any functional effect of four variants exclusive of SCZ patients, the previously mentioned p.A1731S and three singleton variants with high probability for functional effect, according to bioinformatic predictions. These experiments confirmed some degree of functional impairment for each one of the variants. However, there were 10 variants exclusive of SCZ patients and 13 exclusive of controls. Without a comparison of the functionality of variants in cases versus controls, such as in the study about SHANK2 mutations in autism spectrum disorders by Leblond et al.,6 the functionality of the four tested variants is not an evidence of a role in SCZ susceptibility. Furthermore, assuming that each sample presents just one rare variant, 2.70% of patients and 2.73% of controls are carriers of rare variants not shared between cases and controls. The variants responsible for the significant association with schizophrenia are those shared by patients and controls, carried by 4.16% of patients versus 1.21% of controls. Functionality has been detected previously for one of these variants, p.R818H, but there are no data for the remaining two (Supplementary Table 8 of Peykov et al.). In addition, p.R569H, the variant excluded from the genetic analysis by the above-mentioned bias, is also functional. Overall, there are 20 schizophrenic patients and 22 controls that are carriers of functional variants (P=0.52, two-sided Fisher’s exact test). Therefore, data from Peykov et al. did not detect an excess of rare functional variants in cases.
In summary, SHANK2 is an attractive functional candidate gene for involvement in SCZ susceptibility but the arguments present in this letter cast doubt on the Peykov et al.’s claim for a causative role of rare SHANK2 variants in SCZ. On the contrary, this work is an example of the challenges in the analysis of rare variants. In order to draw a conclusion about the role of SHANK2 in SCZ susceptibility, there is a need for additional sequencing data and a better classification of variants as functional versus non-functional, perhaps by the experimental approach of Peykov et al., probably in combination with more powerful tests for the cumulative analysis of rare variants under different scenarios.7, 8
Peykov S, Berkel S, Schoen M, Weiss K, Degenhardt F, Strohmaier J et al. Mol Psychiatry 2015; 10.1038/mp.2014.172.
Lemire M . Nat Genet 2011; 43: 391–392.
Pearson RD . Nat Genet 2011; 43: 392–393.
Berkel S, Marshall CR, Weiss B, Howe J, Roeth R, Moog U et al Nat Genet 2010; 42: 489–491.
1000 Genomes Project Consortium Nature 2012; 491: 56–65.
Leblond CS, Heinrich J, Delorme R, Proepper C, Betancur C, Huguet G et al PLoS Genet 2012; 8: e1002521.
Ionita-Laza I, Lee S, Makarov V, Buxbaum JD, Lin X . Am J Hum Genet 2013; 92: 841–853.
Lee S, Emond MJ, Bamshad MJ, Barnes KC, Rieder MJ, Nickerson DA et al Am J Hum Genet. 2012; 91: 224–237.
JC is supported by grants FIS/FEDER PI11/00770 and CP11/00163 from Instituto de Salud Carlos III. I thank the NHLBI GO Exome Sequencing Project and its ongoing studies that produced and provided exome variant calls for comparison: the Lung GO Sequencing Project (HL-102923), the WHI Sequencing Project (HL-102924), the Broad GO Sequencing Project (HL-102925), the Seattle GO Sequencing Project (HL-102926) and the Heart GO Sequencing Project (HL-103010).
The author declares no conflict of interest.
About this article
Cite this article
Costas, J. The role of SHANK2 rare variants in schizophrenia susceptibility. Mol Psychiatry 20, 1486 (2015). https://doi.org/10.1038/mp.2015.119
Activation of the medial preoptic area (MPOA) ameliorates loss of maternal behavior in a Shank2 mouse model for autism
The EMBO Journal (2021)
meQTL and ncRNA functional analyses of 102 GWAS-SNPs associated with depression implicate HACE1 and SHANK2 genes
Clinical Epigenetics (2020)
Application of Human-Induced Pluripotent Stem Cells (hiPSCs) to Study Synaptopathy of Neurodevelopmental Disorders
Developmental Neurobiology (2019)
Shank2 Deletion in Parvalbumin Neurons Leads to Moderate Hyperactivity, Enhanced Self-Grooming and Suppressed Seizure Susceptibility in Mice
Frontiers in Molecular Neuroscience (2018)
JCI Insight (2017)