Imaging Genetics Study For The Volume Of Human Subcortical Brain Structures

AimsSubcortical brain structures are associated with movement,consciousness,emotion and learning.In previous studies,the genetic associations of the volumes of subcortical brain structures were deeply explored using imaging genetics.However,almost all existing studies are based on European population,which is not enough to disclose the genetic mechanisms of brain imaging phenotypes.First,although the sample size is large,the results based on European populations cannot be directly applied to non-European population because of the differences in genetic structure.Therefore,this study used the CHIMGEN project to explore the genetic assocaitions of the subcortical brain structures’ volumes in non-European population for the first time,thus filling the gap of existing big data studies.Second,a growing number of studies have demonstrated the advantages of trans-ethnic genome-wide association analysis in terms of signal validation,meta-analysis,fine-mapping,and phenotypic variance explained.Therefore,we conducted a trans-ethnic analysis of subcortical brain structures’ volumes by comparing and integrating the results of the CHIMGEN project with the results of the previous UKBB and ENIGMA projects.MethodsFor genetic data,we used PLINK for quality control,and used SHAPEIT and IMPUTE2 for imputation.For imaging data,we used Freesurfer for preprocessing and phenotype extraction.By controlling for covariates,we used BGENIE to conduct genome-wide association analysis on genetic data and imaging data.In order to ensure the reliability and generalibility of the GWAS results,we divided the samples into the discovery dataset and the validation dataset by scanners.The discovery dataset was scanned by GE DISCOVERY MR750,while the validation dataset was scanned by other scanners.In addition,we estimated the heritability of subcortical brain structures’ volume and performed association analysis at the gene level to identify genes that influence the volumes of subcortical brain structures.We compared and integrated the GWAS results of subcortical brain structures’ volumes in CHIMGEN with those published by UKBB and ENIGMA.On the one hand,the newly discovered loci and replicated loci were confirmed by direct comparison,and on the other hand,new loci were further identified on the basis of expanded sample size through meta-analysis.In addition,we used different combinations of meta-analyses to compare whether the results of trans-ethnic meta-analysis were significantly better than results of same-ethnic meta-analysis in terms of phenotypic variance explained and fine-mapping.ResultsThrough genome-wide association analysis,we found a total of 15 loci associated with the volumes of subcortical brain structures in the discovery dataset,10 of which could be replicated in the validation dataset.Correlation analysis at the gene level demonstrated that 19 genes were associated with the volumes of subcortical brain structures.The heritability of the subcortical brain structures’ volume ranged from 0.177 to 0.490.The heritability was lower than that reported in earlier twin studies,but was comparable to the heritability estimated in UKBB.By comparing the results of genome-wide association analysis of CHIMGEN,UKBB and ENIGMA,10 associations were found in CHIMGEN and 12 associations in UKBB,while in ENIGMA,with a sample size 4-6 times larger than the former two studies,29 associations were found.Among them,there were 2 consistent associations between CHIMGEN and UKBB,4 consistent associations between CHIMGEN and ENIGMA,and 9 consistent associations between UKBB and ENIGMA.A total of 42 significant associations were found in a trans-ethnic meta-analysis using data from CHIMGEN,UKBB and ENIGMA,and 16 of which had not been reported in previous studies.The GWAS results of hippocampal volume in ENIGMA showed a higher degree of explanation for the hippocampal volume in UKBB(2.42% for the left hippocampus and 2.21% for the right hippocampus)than that in CHIMGEN(1.30%for the left hippocampus and 1.07% for the right hippocampus).The result of meta-analysis between ENIGMA and UKBB would further widen the gap,that is,explanation for the hippocampal volume in UKBB increased significantly(6.14% for the left hippocampus and 5.24% for the right hippocampus),while there was only a slight increase in the explanation of the hippocampal volume in CHIMGEN(1.95%for the left hippocampus and 1.37% for the right hippocampus).However,only the results of meta-analysis between ENIGMA and CHIMGEN could significantly increase the degree of explanation for the hippocampal volume in CHIMGEN(4.92%for the left hippocampus and 3.26% for the right hippocampus).We performed fine-mapping for the significant loci in GWAS results of hippocampal volume in ENIGMA.And we found that there was no significant improvement in fine-mapping when the UKBB data was added to the ENIGMA data(P = 0.1081).While the 99% credible set was significantly narrowed down when the CHIMGEN data was added to the ENIGMA data(P = 0.0051).ConclusionUsing CHIMGEN data,we identified 10 common variants that may influence the volumes of human subcortical brain structures through known developmental pathways such as neuronal development,neuronal morphology,axon guidance,and neuronal regeneration.Gene-level association analysis also identified several genes that affect neuronal differentiation and migration and are associated with neuropsychiatric disorders.Comparison of the GWAS results of subcortical brain structures’ volumes among CHIMGEN,UKBB and ENIGMA demonstrated that the replication rate between the same ethnic groups was significantly higher than the replication rate between different ethnic groups.In addition,trans-ethnic meta-analysis by incorporating the CHIMGEN,UKBB,and ENIGMA yielded new genetic associations through increasing sample size.Besides,by comparing the same-ethnic meta-analysis with trans-ethnic meta-analysis,we found that although an increased sample size within European population will generate more variants and explain a larger proportion of the variance within European population,it will also further exacerbate existing disparities in explained variance between European and non-European population.Finally,trans-ethnic meta-analysis was significantly superior to the same-ethnic meta-analysis in doing fine-mapping.