F23. EXOME-WIDE TANDEM REPEATS CONFER LARGE EFFECTS ON SUBCORTICAL VOLUMES

• Published in: European neuropsychopharmacology 2023-10, Vol.75, p.S232-S232
• Main Authors: Abbatangelo, Cristina, Adler, Nina, Moo-Choy, Ashley, Panoyan, Mary Anne, Shi, Yuxin, Parra, Esteban, Polimanti, Renato, Hu, Pingzhao, Wendt, Frank
• Format: Article
• Language: English
• ISSN: 0924-977X; 1873-7862
• DOI: 10.1016/j.euroneuro.2023.08.411

Subcortical structures of the human brain play important roles in cognitive, affective, and social functions. Convergent evidence from genome-wide association studies (GWAS) and brain imaging support the role of subcortical abnormalities in many mental illnesses such as depression and schizophrenia. Biological processes/mechanisms linking subcortical volumes to psychiatric conditions are challenging to infer due to the non-coding region localization of GWAS-identified SNPs. Exome-wide tandem repetitive elements (TREs) can model relatively large and independent effects on subcortical volumes that, in turn, influence psychiatric conditions with clear biological influences. UK Biobank (UKB) participants were subjected to T1-weighted structural imaging with a Siemens Skyra 3T. T1-weighted structural imaging was performed over a 5-minute timeframe with 1 × 1 × 1 mm resolution and a 208 × 256 × 256 field-of-view matrix. Subcortical volumes of the accumbens, amygdala, hippocampus, pallidum, putamen, and thalamus were modeled from T1 data using FIRST (FMRIB's Integrated Registration and Segmentation Tool). We summed left and right hemisphere volumes per region. UKB BAM files aligned to the hg38 reference genome were subjected to genotyping of autosomal TREs with GangSTR v2.5.0. Genotypes for European ancestry (N=21,538) participants were converted to a locus-level burden by summing alleles at each TRE. Generalized linear models were used to regress TRE locus burden on subcortical volumes using age, sex, sex × age, age2, sex × age2, and the first 10 within-ancestry principal components as covariates. Multiple testing correction was applied with a Benjamini-Hochberg false discovery rate of 5%. Fine-mapping was performed with FINEMAP v1.4.2 using SNPs in a 1Mb region. There were 18 TREs associated with subcortical volumes (FDR < 5%), 13 of which associated with accumbens volume. Five TRE associations survived a traditional genome-wide significance threshold; these include accumbens-associated loci: ABLIM2-[ACCC]N (intronic; β=6.53, P=1.92 × 10-8, fine-mapping posterior inclusion probability (PIP)=0.883), NTN1-[GCGG]N (3’-UTR; β=5.93, P=8.16 × 10-9, PIP=0.997), SPTBN4-[GGGGC]N (intronic; β=4.37, P=1.50 × 10-8, PIP=0.987), and ZMIZ2-[GTGGG]N (intronic; β=8.79, P=6.25 × 10-9, PIP=0.998) and thalamus-associated locus SLC39A4-[CAG]N (exonic; encodes ZIP4; β=-1559.2, P=2.42 × 10-8, PIP=0.993). Using gene-based results from the GWAS Atlas, ABLIM2 was enriched for associatio