An ENIGMA Consortium Genome-Wide Association Study of Brain Age
Poster No:
1172
Submission Type:
Abstract Submission
Authors:
Vilte Baltramonaityte1, Constantinos Constantinides1, Marlene Staginnus1, Neda Jahanshad2, Paul Thompson2, James Cole3, Sarah Medland4, Danai Dima5, Esther Walton1, ENIGMA consortium6
Institutions:
1University of Bath, Bath, United Kingdom, 2Imaging Genetics Center, Keck School of Medicine of University of Southern California, Los Angeles, USA, 3University College London, London, United Kingdom, 4QIMR Berghofer Medical Research Institute, Brisbane, Australia, 5City University London, London, United Kingdom, 6Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute, Marina del Rey, CA, CA
First Author:
Co-Author(s):
Neda Jahanshad, PhD
Imaging Genetics Center, Keck School of Medicine of University of Southern California
Los Angeles, USA
Imaging Genetics Center, Keck School of Medicine of University of Southern California
Los Angeles, USA
Paul Thompson, PhD
Imaging Genetics Center, Keck School of Medicine of University of Southern California
Los Angeles, USA
Imaging Genetics Center, Keck School of Medicine of University of Southern California
Los Angeles, USA
ENIGMA consortium
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute
Marina del Rey, CA, CA
Imaging Genetics Center, Mark and Mary Stevens Neuroimaging and Informatics Institute
Marina del Rey, CA, CA
Introduction:
Deviations from a typical ageing trajectory are an important risk factor for poor health outcomes.1 Brain-predicted age difference (PAD) – conceptualised as the difference between chronological and brain predicted age – is one such measure of deviation from healthy ageing and has been linked to over 40 traits.2 While brain-PAD is generally thought to be heritable, specific genetic loci that influence these deviations are still largely unknown. Three recent genome-wide association studies (GWASs) on brain-PAD in UK Biobank (n up to 28,104; age range: 40 to 84) identified a small number of associated genetic variants.3-5 This small number might be explained by the narrow age range and moderate sample size used in these studies. Larger samples covering the complete adult lifespan are needed to elucidate the genes implicated in brain-PAD, their impact on other biological systems in the brain and peripheral tissues, and the causal relationship between PAD and mental health.
Methods:
Brain-PAD was derived using a ridge regression model with 77 FreeSurfer-derived structural brain imaging features of surface area, cortical thickness and subcortical volume as an input in a total of n=47,167 participants from 28 datasets within the ENIGMA consortium.6 For a subset of these (n=34,112), we carried out a genome-wide association meta-analysis (GWAS) of brain-PAD. Additive effects of genetic variants on brain-PAD were tested, adjusting for age, age2, sex, total intracranial volume, genetic ancestry, imaging covariates (e.g. multiple scanners) and disease status (for case-control studies). We applied linear (mixed) models using BOLT-LMM, RareMetalWorker or PLINK2. Preliminary results were meta-analysed in METAL, weighing each cohort according to sample size.
Results:
A total of n=47,167 participants were included in the phenotypic analysis (age range 18-75 years; 52.8% females). Brain age was predicted with meanunweighted absolute error of 9.58 years (range 4.67-21.29). For the subset of datasets included in the GWAS, the meanunweighted absolute error was slightly larger (14.25 years; range 6.30-21.29; age-bias correctedunweighted=9.08). Fixed effect meta-analysis using METAL identified 66 genome-wide significant variants associated with brain-PAD at P = 5 x 10-8 (Figure 1). Three of these variants (on chromosomes 2, 15 and 16) were independent using r2 = 0.1 and 500 kb window size. Two out of three variants identified had been previously implicated in brain-related phenotypes. SNP-based heritability was estimated at 0.1923 (SE=0.0167).
·Figure 1. Manhattan plot for METAL meta-analysis of Brain-PAD. Red line: P=5x10-8, blue line: P=5x10-5. Brain-PAD=brain predicted age difference (calculated as predicted minus chronological age).
Conclusions:
Our findings indicate that brain age deviations in adulthood might be moderately heritable. Genetic loci overlapped partially with previous studies using overlapping data (e.g., UK Biobank), but different brain age estimation methods, suggesting a degree of consistency across methods.3,7 Identifying the underlying genetic loci can help to shed light on the causal risk factors involved in brain aging, aiding in the prevention and treatment of age-related poor health outcomes, such as schizophrenia, Alzheimer's disease, and other cognitive impairments.
Genetics:
Genetic Association Studies 2
Lifespan Development:
Aging 1
Keywords:
Other - Brain age
1|2Indicates the priority used for review
Provide references using author date format
2. Kolbeinsson, A. (2020), 'Accelerated MRI-predicted brain ageing and its associations with cardiometabolic and brain disorders'. Scientific Reports 10, 19940.
3. Smith, S. M. (2020), 'Brain aging comprises many modes of structural and functional change with distinct genetic and biophysical associations', eLife 9, e52677.
4. Jonsson, B. A. (2019), 'Brain age prediction using deep learning uncovers associated sequence variants', Nature Communications 10, 5409.
5. Leonardsen, E. H. (2023), 'Genetic architecture of brain age and its causal relations with brain and mental disorders', Molecular Psychiatry 28, 3111–3120.
6. Han, L. K. M. (2021), 'Brain aging in major depressive disorder: results from the ENIGMA major depressive disorder working group', Molecular Psychiatry 26, 5124–5139.
7. Kim, J. (2023), 'Investigation of genetic variants and causal biomarkers associated with brain aging', Scientific Reports 13, 1526.