Phenome-wide association studies between SERINC2 and neuropsychiatric disorders

SERINC2 has been associated with alcoholism, bipolar disorder and autism, but the comparability and specificity issues of the findings remain unaddressed. The present study aimed to comprehensively analyze various neuropsychiatric disorders pinpoint the most reliable conditions predisposed by SERINC2.

A total of 2,187 imputed SNPs across SERINC2 were examined in 1,167,439 subjects from 72 independent cohorts with 18 different neuropsychiatric disorders. SNP-disease associations were tested and then meta-analyzed, followed by FDR correction, to identify significant disease-risk SNPs. Finally, functional studies on the differential SERINC2 mRNA expression in brains and the potential regulatory effects of disease-risk alleles on SERINC2 mRNA expression, gray matter volumes (GMVs) of subcortical structures, cortical surface area (SA) and average thickness (TH) were conducted.

In European descent, alcoholism was most significantly associated with SERINC2 variants (245 SNPs with 5.5×10^-8≤p ≤ 0.049 and 4.9×10^-5≤q ≤ 0.034) that were largely shared across cocaine dependence, marijuana dependence, nicotine dependence, polysubstance dependence, schizophrenia, OCD, and autism (8.2×10^-8≤p ≤ 0.050 and 1.9×10^-5≤q ≤ 0.049); in Chinese population, bipolar disorder was also significantly associated with SERINC2 variants (10 SNPs: 1.3×10^-4≤p ≤ 4.7×10^-4 and 0.025≤q ≤ 0.031). Furthermore, the disease-risk alleles had highly similar regulatory effects on mRNA expression (8.1×10^-7≤p ≤ 0.046), subcortical GMVs (7.0×10^-4≤p ≤ 0.048) and cortical TH and SA (1.3×10^-3≤p ≤ 0.050) in brains across alcoholism, schizophrenia, OCD and autism. The bipolar disorder-risk alleles had these regulatory effects but with different effect patterns. Finally, SERINC2 mRNA was differentially expressed in several brain regions between alcoholism or schizophrenia and controls.

SERINC2 is primarily linked to substance use disorders, schizophrenia, OCD, autism and bipolar disorder, not only statistically but also biologically.