PC) of Surface Area Measurements in 34 Cortical Regions: PC1 (Panel A) and PC2 (Panel B)
Total Page:16
File Type:pdf, Size:1020Kb
Supplementary Figures Figure E1. Loadings for the first two leading principal components (PC) of surface area measurements in 34 cortical regions: PC1 (panel A) and PC2 (panel B). Figure E2. Loadings for the first two leading principal components (PC) of thickness measurements in 34 cortical regions: PC1 (panel A) and PC2 (panel B). Figure E3. Manhattan plots for the meta-analyses of GWAS for surface area PC1 (panel A) and PC2 (panel B) scores in CHARGE and UKBB cohorts. Figure E4. Forest plot of the top PC2 SNP rs73313052. Figure E5. Manhattan plots for the meta-analyses of GWAS thickness PC1 (panel A) and PC2 (panel B) scores in CHARGE and UKBB cohorts. Figure E6. DAAM1 co-localizes with mitochondrial marker ATP5A in the primary visual cortical plate. A. Figure E7. Associations between lateral geniculate nucleus (LGN) volume and surface area PC2 scores in 694 SYS major-homozygotes (GG) (panel A) and in the 206 carriers of minor allele of rs73313052 (GA or AA) (panel B). Figure E8. Genetic correlation between schizophrenia vs. surface area PC1 and PC2. Figure E9. Genetic correlation between ADHD vs. surface area PC1 and PC2. Figure E10. Genetic correlation between autism spectrum disease (ASD) vs. surface area PC1 and PC2. Figure E11. Genetic correlation between bipolar disorder (BD) vs. surface area PC1 and PC2. Figure E12. Genetic correlation between major depressive disorder (MDD) vs. surface area PC1 and PC2. Figure E13. Genetic correlation between neuroticism vs. surface area PC1 and PC2. Figure E14. Genetic correlation between extraversion vs. surface area PC1 and PC2. Figure E15. Genetic correlation between neo-agreeableness vs. surface area PC1 and PC2. Figure E16. Genetic correlation between neo-agreeableness vs. surface area PC1 and PC2. Figure E17. Genetic correlation between neo-openness vs. surface area PC1 and PC2. Figure E18. Genetic correlation between global cognitive function vs. surface area PC1 and PC2. Figure E19. Genetic correlation between educational attainment vs. surface area PC1 and PC2. Figure E20. Genetic correlation between anorexia nervosa vs. surface area PC1 and PC2. Figure E1. Loadings for the first two leading principal components (PC) of surface area measurements in 34 cortical regions: PC1 (panel A) and PC2 (panel B). Each cohort estimated the surface area of the 34 cortical regions (left and right hemispheres summed), using FreeSurfer and carried out unrotated principal component analyses to obtain PC loading values for the first two leading PCs, which are indicated by colored lines: 0-indigo (ARIC), 1-light-blue (ASPSFam), 2-green (BIL&GIN), 3-yellow (CHS), 4-red (FHS), 5-purple (LIFE-Adult), 6-grey-blue (RS), 7-olive-green (SYS-Adult), and 8-mint-green (SYS- Adolescent). Black thicker lines indicate the median loading values across all the cohorts. The median loading values were then used to derive the ‘general’ PC score for each individual and later used as the response variable in the meta-GWAS analyses. 0 ARIC 2 BIL&GIN 4 FHS 6 RS 8 SYS cohort 1 ASPSFam 3 CHS 5 LIFE−Adult 7 SPS A 1.00 4 4 4 4 4 231● 4 3● 4 5 4 83 4 28 2 4 4 4 24 760 1 1 4 25● 6 2 ● 3 0 4 3 ● 4 31 3● 4 2 2 16 61 2 2 51 3 3 2● 5 4 031 4 4 3● 73 ● 8 70 ● 1 56 7 4 ● 6 7 4 2 8 1 0 0 4 2 41 3 3● 05 27 2 3● 1 60 8 5 60 4 580 2578 756 7 ● 1 32● 0 1 7 2 ● 15 6 16 2● 01 5 4 5 2 0.75 15 67 76 1 2 7 3 8 87 6 8 5 ● 0 1 2 170 12 1 0 32 8 50 2 506 85 6 6 05● 86 4 281● 2 8● 1 7 30 1 8 743 8 ● 5 5 ● 7 3 3 03 5 3 8 7 87 1 06 7 2 4 8 0 8 6 3● 2 3 8 7 1 1 7● 5 3 8 465 7 4 8 0817 34 ● 3 5 ● ● 0 8 1 6 ● 7 0.50 067 5 3 4 0● 3 80 1 08 0 5 752 6 6 7265 5 PC1 PC1 13 0 6 0 2 6 2 3 ● ● 7 ● 5 8 75 8 25 6034 0 ● 7 7 8 167 68 6 8 0.25 6 5 8 6 0 0 ARIC 2 BIL&GIN 4 FHS 6 RS 8 SYS cohort 1 ASPSFam 3 CHS 5 LIFE−Adult 7 SPS 0.00 0 ARIC 2 BIL&GIN 4 FHS 6 RS 8 SYS frontalpole rostralmiddlefrontal parsorbitalis medialorbitofrontal rostralanteriorcingulate cohortlateralorbitofrontal parstriangularis superiorfrontal caudalanteriorcingulate parsopercularis caudalmiddlefrontal temporalpole insula entorhinal precentral superiortemporal posteriorcingulate postcentral transversetemporal middletemporal paracentral parahippocampal inferiortemporal supramarginal fusiform bankssts isthmuscingulate precuneus superiorparietal lingual inferiorparietal cuneus pericalcarine lateraloccipital 1.0 1 ASPSFam 3 CHS 5 LIFE−Adult 7 SPS 6 B 6 6 07 7 3● 0 85 650 1.00 ● 0.5 1 1 2 3● 60 825 7● 8 4 3 305 71 056 1 ● 4 1 2 5 8● 8 5 6 0 68● 2 82 7352 5 8 5 1 ● 3 ● 0 241 8 5 3 50 8● 7 ● 5 0 2 1 7 ● 6 6 8 ● 8 5 8 ● ● 2 3 26 8● 305 6 2 ● 4 4 0 7● 7 7 ● 6 2 ● 8 8 3 4 4 ● ● 4 4 35 0 7 66 17 742 60 45 6 4 1 7 ● 78 38 4 3 1273 5 0 27 3 802 21 6 ● 2 6 508 1 8 7 75 1 0 6 7 57 7● ● 75 0 4 ● 4 ● 20 4 8 0 4 44 4 2 7 ● ● 3 1 1 06 3 0 4 1 6 5 0 180 ● 6 ● 1 2 4 7 6 13 8 1 63 4 4 ● 6 6 3 78 4 6 28 ● 6582 563 25 5728 42 6 03 0 52● 4 0 2 2● 2 8 1 0 2 7 4 53 3 2 7 5 2 58 7 36 5 4 ● 30 63 74 023 5 4 16 ● 8 705 3 70 4● 8 382 1 1 1 5● 02 42 0 0 18 328 2 74 748 4 35 84 6207 ● 6 87 5 2 1 0 80 1 6 6 37 0.75 3 5 7 1 34 67 2● 0.0 3 1 0 ● 4 2 1 4 70 ● 4 6 6 3 3 65 8 6302 13 13 ● 28 3 8 1 1 560 ● 15 6 8 7 3 6 1 6 ● 1 4 7 5 1 74 1 3 6 23 5 ● 47 ● 5 ● 2 7 6 ● 0 ● 8 ● 7 3 5 0 0 PC2 1 14 1 0 3 7 5 6● ● PC2 35 3 06 3 6 1 815 15 7 1 23 0571 42 1348 536 80 8 4 7 12 ● 1 ● 3 0 56 72 52 2 7 3● 5 5 ● 6 7 023 0 028 8 6 0 4870 3 4 6 658 685 41 48 3 574 0721 6● 1 6 247 1 ● 4 7 1 3 2 82 4 4 5 6 3 0 6 0 72 ● 4 8 8 6 4 ● 71 8 ● 7420 42 0257 41 28 85 1 30 7 2750 740 6 8 8 3● 1306 1 ● 7 0● 3 2 1 5 5 6 052 573 8 −0.506 3 8 41 45 0.50 ● 4 82 48 154 8 1 PC1 8 3 3 1 0.25 −1.0 frontalpole rostralmiddlefrontal parsorbitalis medialorbitofrontal rostralanteriorcingulate lateralorbitofrontal parstriangularis superiorfrontal caudalanteriorcingulate parsopercularis caudalmiddlefrontal temporalpole insula entorhinal precentral superiortemporal posteriorcingulate postcentral transversetemporal middletemporal paracentral parahippocampal inferiortemporal supramarginal fusiform bankssts isthmuscingulate precuneus superiorparietal lingual inferiorparietal cuneus pericalcarine lateraloccipital 0.00 frontalpole rostralmiddlefrontal parsorbitalis medialorbitofrontal rostralanteriorcingulate lateralorbitofrontal parstriangularis superiorfrontal caudalanteriorcingulate parsopercularis caudalmiddlefrontal temporalpole insula entorhinal precentral superiortemporal posteriorcingulate postcentral transversetemporal middletemporal paracentral parahippocampal inferiortemporal supramarginal fusiform bankssts isthmuscingulate precuneus superiorparietal lingual inferiorparietal cuneus pericalcarine lateraloccipital Figure E2. Loadings for the first two leading principal components (PC) of thickness measurements in 34 cortical regions: PC1 (panel A) and PC2 (panel B). Each cohort estimated the thickness of the 34 cortical regions (left and right hemispheres averaged), using FreeSurfer and carried out unrotated principal component analyses to obtain PC loading values for the first two leading PCs, which are indicated by colored lines: 0-indigo (ARIC), 1-light-blue (ASPSFam), 2-green (BIL&GIN), 3-yellow (CHS), 4-red (FHS), 5-purple (LIFE-Adult), 6-grey-blue (RS), 7-olive-green (SYS-Adult), and 8-mint-green (SYS- Adolescent). Black thicker lines indicate the median loading values across all the cohorts. The median loading values were then used to derive the ‘general’ PC score for each individual and later used as the response variable in the meta-GWAS analyses. CRHR1 A MAPT STH KANSL1 PC1 B DAAM1 PC2 Figure E3. Manhattan plots for the meta-analyses of GWAS for surface area PC1 (panel A) and PC2 (panel B) scores in CHARGE and UKBB cohorts. The GWAS meta- analyses were carried out as follows. Using the median PC loadings across the CHARGE consortium cohorts (Figure E1), each cohort derived the ‘general’ PC scores to be used as phenotypes in the genome-wide association tests in order to ensure ‘homogeneity’ in phenotype derivation. All the association tests were adjusted for age, sex and other cohort- specific confounding variables, such as study site and/or family structure. The cohort-specific GWAS results were then examined for quality-control with Easy QC software1, and meta- 2 analyzed with METAL using fixed effects models.