The hippocampus is essential for memory, cognition, and neuropsychiatric disorders. Recently, Tianjin Medical University Yu Chunshui, Ye Zhaoxiang and Wang Meiyun of Zhengzhou University collaborated on cross-population genome-wide association analysis, and found that 44 hippocampal traits were associated with 339 variants, including 23 new associations. Common genetic variants affect the hippocampal characteristics of various populations, but there are also population-specific associations. Cross-population analysis improves granular mapping and polygenic fraction prediction.
Their results, "Cross-ancestry genome-wide association meta-analyses of hippocampal and subfield volumes," were recently published in the journal Nature Genetics.
The hippocampus is essential for memory and cognitive function and is affected by a variety of neuropsychiatric disorders. The size of the hippocampus is highly genetic, but the genetic association between populations varies greatly, and diversified GWAS studies are required. The Chinese Imaging Genetics Project collected data from Han Chinese participants to reveal the genetic structure of hippocampal volume through cross-ancestral GWAS meta-analysis. Research aims to discover new associations, improve the accuracy of detailed mapping, and assess predictive capabilities. At the same time, the research also focuses on the functional and clinical relevance of the found genetic variation.
Subject
In this study, the volume of 44 hippocampal and sub-regional structures was symmetrically distributed in the left and right cerebral hemispheres using cross-population GWAS meta-analysis (Figure 1a). The authors combined GWAS data from EAS and EUR participants. EAS-GWAS data for hippocampal volume were from Han Chinese participants (7,009 people), autosomal EUR-GWAS data were from UK Biobank (31,968 people), and EUR-GWAS data for hippocampal sub-volume were from UKBB (31,968 people). EUR-GWAS data for PAR (pseudo-autosomal regions) X chromosome variants were from 31,943 UKBB participants and data for non-PAR variants were from 31,954 participants.
Figure 1.Average activation of each sentiment category in two videos (model coefficient or beta)
GENETIC ASSOCIATION OF CHIMGEN
EAS-GWAS studies of 44 hippocampal volume traits were performed in 7,009 participants, analyzed using 6,830,145 SNPs, and 25 significant associations were found (P<5×10−8). Of these associations, three survived P < 1.13 × 10−9 when an additional 44 hippocampal features were corrected, including rs10901817 (chr10) for left hippocampal volume, rs10901817 for left hippocampal tail volume (P = 2.81 × 10−10), rs199840783 (chr10) for right hippocampal tail volume, and rs6496265 (chr15) for left hippocampal volume (P = 6.58 × 10−10). Further GWAS was performed for 44 traits, 181,603 X chromosome SNPs, and no genome-wide significant associations were found.
Cross-population genetic associations
The authors analysed a total of 4,901,971 autosomal variants in the GWAS study at CHIMGEN, UKBB and ENIGMA. A cross-population GWAS meta-analysis of the left and right hippocampal volumes of 7009 EAS participants from CHIMGEN and 58,782 EUR participants from UKBB and ENIGMA found 97 significant associations. Of the 5,110,460 autosomal variants shared, meta-analysis of 42 hippocampal sub-regions of 7,009 CHIMGEN and 31,968 UKBB participants found 508 significant associations. A total of 605 significant associations involved 518 trait association sites, of which 137 were previously unreported. Combining the association signals and sites of 44 hippocampal features, the authors found 126 independent association signals (102 sites), of which 27 association signals and 26 loci were not previously reported. Twenty-three novel associations were also discovered, involving 303 traits-related sites (Figure 1B).
Population sharing and population-specific genetic associations
Population sharing and specific genetic associations of these brain structures between Europeans and East Asians remain unknown due to the lack of genetic studies of hippocampal and self-extractive region volumes in non-European populations. Genetic variation analysis was carried out on Europeans and East Asians, and it was found that 84.51% of the associations were in the same direction, and 72.01% of the variants were not heterogeneous. However, a 5.71% specific genetic association was also found, associated with race-specific genetic effects (Figure 2).
Figure 2. Population sharing and population-specific genetic associations of hippocampal and subregional volumes
Fine mapping across populations
In order to further investigate the causal relationship between hippocampus and subregion volume, fine mapping was performed on 303 traits-related sites. Cross-population fine mapping found important loci, such as rs7315280 and rs7966895, associated with hippocampal volume and neurodevelopment. Cross-population analysis provides more accurate and detailed results than single-population analysis, increasing the number of fine mappings (Figure 3).
Figure 3. Cross-population analysis improves fine mapping resolution
Transferability of PGS (polygenic score) for cross-population meta-analysis
Current large-scale GWAS studies have focused on European individuals, resulting in PGS constructed from these individuals performing poorly in predicting the characteristics of non-European individuals. In this study, the applicability of PGS constructed based on different population combinations in predicting hippocampal characteristics in Asian populations was systematically evaluated. The results showed that the prediction performance of PGS in cross-population analysis was better than that of Asian-specific and European-specific analysis, indicating that the predictive ability of underrepresented population characteristics can be improved by adding moderate Asian samples (Figure 4).
Figure 4. PGS constructed using different protocols to predict hippocampal volume traits
Functional annotation of genetic variation
To investigate the function of genetic variation associated with hippocampal and subregion volume, the authors functionally annotated 1000 SNPs with the highest PP values using FUMA. The results showed that these SNPs were mainly located in the intronic and intergenic regions. In addition, mutations associated with the hippocampus, such as MADD and APOE, were identified, as well as genes associated with neurodevelopment and immunity. Colocalization analysis also revealed sites associated with hippocampal and subfield volume, and identified genes associated with Wnt signaling pathways and neuronal differentiation (Figure 5).
Figure 5.Functional annotation of genetic variants associated with hippocampal and subregion volume
Genetic colocalization with brain-associated phenotypes
The Coloc approach found genetic colocalization of hippocampal volume and brain-associated phenotypes, including neuropsychiatric disorders. Colocalization sites include genes such as APOE and SLC4A10, which are associated with traits such as cognitive, affective, and neurotic traits (Figure 5d, e, f).
conclusion
The authors performed a cross-population GWAS meta-analysis to discover new genetic associations in hippocampus and subregion volume and provide biological insights into neuropsychiatric disorders. In addition, the authors demonstrated that small sample genetic data from underrepresented populations can improve genetic risk prediction. However, due to the lack of independently replicated datasets and other limitations, we still need to interpret these results with caution.
Original link:
https://doi.org/10.1038/s41588-023-01425-8
bibliography
Liu N, Zhang L, Tian T, et al. Cross-ancestry genome-wide association meta-analyses of hippocampal and subfield volumes [published online ahead of print, 2023 Jun 19]. Nat Genet.2023;10.1038/s41588-023-01425-8. doi:10.1038/s41588-023-01425-8
Compiled by Ayden (brainnews creative team)
Proofreader: Simon (Brainnews Editorial Office)