PC) of Surface Area Measurements in 34 Cortical Regions: PC1 (Panel A) and PC2 (Panel B)

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 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. The y-axis indicates the –log10 P values of the SNPs in the meta-analyses. The red dotted line indicates the genome-wide significance level of 5.0×10-8. The most significant loci are highlighted with black-dashed eclipses. The labelled genes within the PC1-GWAS-identified locus on chromosome 17 have been shown to be associated with intracranial volume previously. The DAAM1 gene within the PC2- GWAS-identified locus on chromosome 14 contains multiple SNPs in moderate LD (r2 > 0.4) with the top SNP rs73313052.

1Winkler, T. W. et al. Quality control and conduct of genome-wide association meta- analyses. Nat. Protoc. 9, 1192–1212 (2014).

2Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26, 2190–2191 (2010).

Figure E4. Forest plot of the top PC2 SNP rs73313052. Forest plots show the effect at each of the contributing cohort to the meta-analysis. The cohort-specific effect size estimates are indicated by the open circles, and the meta-analysis effect size estimate, by an open diamond; and their 95% confidence intervals (CI) are shown by the horizontal lines. The size of each point and width of CI lines are proportional to the sample size.

A RAB11FIP4 PC1

B PC2

Figure E5. Manhattan plots for the meta-analyses of GWAS thickness 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 E2), 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-analyzed 2 with METAL using fixed effects models. The y-axis indicates the –log10 P values of the SNPs in the meta-analyses. The red dotted line indicates the genome-wide significance level of 5.0×10-8. The most significant loci are highlighted with black-dashed eclipses. The labelled gene for PC1 is the mapped coding gene to the GWAS-significant locus (rs9907422; p=2.4×10-8). For PC2, there was no protein coding genes mapped to the GWAS-significant locus (rs7196814; p=1.7×10-9).

1Winkler, T. W. et al. Quality control and conduct of genome-wide association meta- analyses. Nat. Protoc. 9, 1192–1212 (2014).

2Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26, 2190–2191 (2010).

A B

30 μm

Fig E6. DAAM1 co-localizes with mitochondrial marker ATP5A in the primary visual cortical plate. A. Schematic representation of the medial view of a midfetal (22 post- conception weeks) human brain; vertical line indicates the level at which coronal sectioning was performed for immunofluorescent experiments. B. Co-localization of DAAM1 and ATP5A in the human midfetal primary visual cortical plate. DAAM1 (green) can be observed in the cytoplasm of cells, whereas ATP5A (red) is localized in mitochondria, which are enriched around the nucleus (labeled with DAPI in blue) of cells. Scale bar represents 30 µm.

A. rs73313052 – 694 GG carriers B. rs73313052 – 206 GA or AA carriers SA-PC2 score (adjusted)

r = 0.13 (P = 0.0006) r = -0.06 (P = 0.37)

LGN volume (adjusted)

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). The x- and y-axes indicate, respectively, LGN volume and surface area PC2 score, both adjusted for intracranial volume. The correlation coefficients between LGN and PC2 scores were 0.13 (P=0.005) and -0.06 (P=0.37), respectively, in GG-carriers and in minor-allele carriers, and t-test indicates that rs73313052-by-LGN interaction effect on PC2-scores is significant (t=-1.97; P <0.05) in 900 pooled individuals. The LGN volume was estimated as follows. First, the Jülich Histological atlas was used to obtain a cytoarchitectonic probabilistic map of the lateral geniculate nucleus (LGN) within MNI space1; the LGN mask was defined as voxels with a probability of P > 50%. Second, the thresholded LGN mask was non-linearly registered to a group template space (SYS808) using Advanced Normalization Tools (ANTs)2. The SYS808 template was created through a series of non-linear registrations of 808 T1-weighted images from the Saguenay Youth Study (SYS)3. Prior to registering LGN masks in SYS808 space to native subject space, individual T1-weighted scans are preprocessed by correcting for inhomogeneities using N3 (nonparametric non-uniformity normalization)4 and skull stripping (Brain Extraction Tool)5. Third, the LGN mask from SYS808 template space was warped into individual space using diffeomorphic warping in ANTs2, and mask volume was extracted using FSL toolkit6. ANTs diffeomorphic registration protocol was adapted from a previously reported methodology for extracting LGN Jacobian determinants7. Left LGN volumes were extracted for 925 participants from the SYS dataset (448 males, 494 females, average age 15.1 years)3. References

1. Bürgel, U. et al. White matter fiber tracts of the human brain: three-dimensional mapping at microscopic resolution, topography and intersubject variability. Neuroimage 29, 1092–1105 (2006). 2. Avants, B. B., Tustison, N. & Song, G. Advanced normalization tools (ANTS). Insight J 2, 1–35 (2009). 3. Paus, T. et al. Saguenay Youth Study: A multi-generational approach to studying virtual trajectories of the brain and cardio-metabolic health. Dev. Cogn. Neurosci. 11, 129– 144 (2015). 4. Sled, J. G., Zijdenbos, A. P. & Evans, A. C. A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans. Med. Imaging 17, 87–97 (1998). 5. Smith, S. M. Fast robust automated brain extraction. Hum. Brain Mapp. 17, 143–155 (2002). 6. Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W. & Smith, S. M. Fsl. Neuroimage 62, 782–790 (2012). 7. Aguirre, G. K. et al. Patterns of individual variation in visual pathway structure and function in the sighted and blind. PloS One 11, e0164677 (2016).

PPP1R21 FOXN2 FANCL ANKRD44 FOXO3 A SLC39A8 HSPD1 BANK1 HSPE1 ELF2 HSPE1-MOB4 MOB4 HMGA2 RFTN2 C10orf32-ASMT MYCL COQ10B NEK4 AS3MT CRHR1 NUCB2 ST7L SF3B1 PPP2R3A WNT3 RPEL1 PLEKHA7 FOXN2 PLCL1 MAPT OPCML BOLL PC1

JKAMP CCDC175 L3HYPDH B GPR135 PPP1R13B ELOVL7 ERCC8 XRCC3 AL049840.1 NDUFAF2 SMIM15 ZFYVE21 MARK3 BAG5 DPP4 AC008498.1 ANKRD44 LIN28B NDFIP2 KLC1 DEPDC1B PLCL1 RBM26 RP11-73M18.2 NEUROD2 RFTN2 APOPT1 PPP1R1B EVI5 STARD3 IGSF9B MED1 NCAPD3 SLC6A11 PC2

Figure E8. Genetic correlation between schizophrenia vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for schizophrenia were from meta-analysis2 of 77,096 individuals in Psychiatric Genomics Consortium (PMID: 25056061).

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Ripke S, Neale BM, Corvin A, Walters JT, Farh KH, Holmans PA, Lee P, Bulik-Sullivan B, Collier DA, Huang H, Pers TH. Biological insights from 108 schizophrenia-associated genetic loci. Nature. 2014 Jul;511(7510):421.

A RNF10 POP5 GATC SRSF9 DYNLL1

CRIM1 PC1

B FBXL17

SLC6A11 PC2

Figure E9. Genetic correlation between ADHD vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for schizophrenia were from meta-analysis2 of 77,096 individuals in Psychiatric Genomics Consortium (PMID: 25056061).

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Demontis D, Walters RK, Martin J, Mattheisen M, Als TD, Agerbo E, Belliveau R, Bybjerg- Grauholm J, Bækved-Hansen M, Cerrato F, Chambert K. Discovery of the first genome-wide significant risk loci for ADHD. BioRxiv. 2017 Jan 1:145581.

CAMSAP3 A

CRHR1 WNT3 MAPT PC1

R3HCC1

SLC6A11 PC2

Figure E10. Genetic correlation between autism spectrum disease (ASD) vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y- axis shows the unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests1. The red dotted line indicates the genome-wide significance level of 5.0E- 08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for ASD are from the first PGC Autism meta- analysis of 5,305 ASD-diagnosed cases and 5,305 psuedo controls of European descent (https://www.med.unc.edu/pgc/files/resultfiles/PGCASDEuro_Mar2015.readme.pdf).

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

A RETSAT ELMOD3 ZBTB38 MCTP1

NCAN PC1

B HDAC5 UPK3B MSTO1 SSBP2 ATXN7L3 ZC3HC1 GON4L CDH13 UBTF PC2

Figure E11. Genetic correlation between bipolar disorder (BD) vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests1. The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for BD are from a meta-analysis2 of 20,129 BD-diagnosed cases and 54,065 controls from Bipolar and Schizophrenia Working Group of the Psychiatric Genomics Consortium.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Ruderfer DM, Ripke S, McQuillin A, Boocock J, Stahl EA, Pavlides JM, Mullins N, Charney AW, Ori AP, Loohuis LM, Domenici E. Genomic Dissection of Bipolar Disorder and Schizophrenia, Including 28 Subphenotypes. Cell. 2018 Jun 14;173(7):1705-15.

A PRELID1P1 PC1

B C10orf32 KLC1 C10orf32-ASMT PC2

Figure E12. Genetic correlation between major depressive disorder (MDD) vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y- axis shows the unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests1. The red dotted line indicates the genome-wide significance level of 5.0E- 08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for major depressive disorder are from a meta- analysis2 a genome-wide association (GWA) meta-analysis based in 135,458 cases and 344,901 controls in PGC.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Trzaskowski M, Wray N, Sullivan P. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depressive disorder. Human Genomics. 2018 Mar 9;12. A PC1

MINOS1- NBL1 B MINOS1 NBL1 PC2

Figure E13. Genetic correlation between neuroticism vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for neuroticism are from the GPC GWAS: Harmonized neuroticism and extraversion summary statistics (GPC2): a meta-analysis of GWAS in 29 discovery cohorts (N=63,661) and 1 replication cohort (N=9,786) for two harmonized personality traits: Neuroticism and Extraversion2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Van Den Berg SM, De Moor MH, McGue M, Pettersson E, Terracciano A, Verweij KJ, Amin N, Derringer J, Esko T, Van Grootheest G, Hansell NK. Harmonization of Neuroticism and Extraversion phenotypes across inventories and cohorts in the Genetics of Personality Consortium: an application of Item Response Theory. Behavior Genetics. 2014 Jul 1;44(4):295-313. A PC1

B PC2

Figure E14. Genetic correlation between extraversion vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for extraversion are from the GPC GWAS: Harmonized neuroticism and extraversion summary statistics (GPC2): a meta-analysis of GWAS in 29 discovery cohorts (N=63,661) and 1 replication cohort (N=9,786) for two harmonized personality traits: Neuroticism and Extraversion2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

A PC1

B PC2

Figure E15. Genetic correlation between neo-agreeableness vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for neo-agreeableness are from the GPC GWAS (GPC1): a meta-analysis of GWA data for personality in 10 discovery samples (17,375 adults) and five in silico replication samples (3,294 adults) of European ancestry2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2De Moor MH, Costa PT, Terracciano A, Krueger RF, De Geus EJ, Toshiko T, Penninx BW, Esko T, Madden PA, Derringer J, Amin N. Meta-analysis of genome-wide association studies for personality. Molecular Psychiatry. 2012 Mar;17(3):337.

A PC1

B PC2

Figure E16. Genetic correlation between neo-agreeableness vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for neo-consciousness are from the GPC GWAS (GPC1): a meta-analysis of GWA data for personality in 10 discovery samples (17,375 adults) and five in silico replication samples (3,294 adults) of European ancestry2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

A PC1

B PC2

Figure E17. Genetic correlation between neo-openness vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for neo-openness are from the GPC GWAS (GPC1): a meta-analysis of GWA data for personality in 10 discovery samples (17,375 adults) and five in silico replication samples (3,294 adults) of European ancestry.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

NCAPG MAP3K14:RNA5SP443 LCORL ARHGAP27:PLEKHM1 A KRT18P63 AC091132.1:LRRC37A4P REST FOXO3 DND1P1:RPS26P8 CENPW CRHR1-IT1:CRHR1 BANK1 TG RAB5B MIR588 SLC39A8 SLC39A4 SUOX MAPT-AS1:SPPL2C AC092989.1 RNU6-200P AF213884.1 MAPT:STH DYSF PPP2R3A TONSL IKZF4 VIMP1 PPP1R16A RPS26 KANSL1:ARL17B TCF7L1 MSL2 NSF:RPS7P11 PRELID1P1 GPT ERBB3 AC093162.5 PCCB RPS4XP9 WNT3:GIP RECQL4 GADD45G HMGA2 TGOLN2 ATP1B3 IGF2BP1 RSPO3 LRRC14 PTCH1 CORO1C Y_RNA TFDP2 OR7E14P BOLL LRRC24 AC116533.1 COX6A1 PLCL1 EPB41L4A ARHGAP39 NUCB2 CREB5 PC1

PDE4D PART1 DEPDC1B KRT8P31 TBC1D5 ELOVL7 B CAMKV ERCC8 ATAD2B ACTBP13 NDUFAF2 EBPL PSMD14 MST1R SMIM15 PTCH1 RBM26 TBR1 CTD-2330K CTC-436P18.1 RBM26-AS1 SLC4A10 9.2 CTC-436P18.3 CENPW NDFIP2-AS1 SSH2 CPEB3 DPP4 CTD-2330K AC011410.1 MIR588 IGF2BP1 MARCH5 NDFIP2 PLCL1 9.3 RNU6-200P LINC01068 CLN3 VIMP1 HIF1AN MON1A AL354741.1 KLC1 APOBR PRELID1P1 RBM6 XRCC3 CCDC101 LIN28A RBM5 REST RPS4XP9 SEMA3F PC2

Figure E18. Genetic correlation between global cognitive function vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests1. The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for global cognitive are from a meta-analysis of genome-wide association data with N=269,867 participants2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Savage JE, et al. Genome-wide association meta-analysis (N=269,867) identifies new genetic and functional links to intelligence. Nature Genetics, forthcoming

MAP3K14 A FOXO3 AKR1E2 ARHGAP27 NCOA7 CASC10 PLEKHM1 CENPW SKIDA1 AC091132.1 MIR588 DND1P1 PHF20L1 MLLT10 RNU6-200P TG DNAJC1 CRHR1-IT1 CRHR1 VIMP1 PPP1R16A CNNM2 PRELID1P1 GPT NT5C2 RAB5B MAPT-AS1 SUOX MAPT BANK1 RPS4XP9 RECQL4 RPEL1 ESR1 IKZF4 KANSL1 SLC39A8 LRRC14 INA RPS26 KANSL1-AS1 AF213884.1 TMEM242 LRRC24 PCGF6 ZBTB38 ERBB3 NSF FNIP2 C8orf82 TAF5 ATP1B3 HMGA2 PC1 ARHGAP39 APBA1 RPS7P11 TFDP2 FOXN2 RNU2-36P FAT3 WNT3 ST7L DLG1 PDE4D BCL11A GADD45G OPCML GIP AKT3 LINC01 RNA5SP30 SALL1 006 IGF2BP1

FBXL20 PDE4D MED1 PART1 TBC1D5 IFT140 CDK12 CAMKV DEPDC1B TMEM204 RNU6-233P KRT8P31 ACTBP13 CLN3 NEUROD2 ELOVL7 B MST1R APOBR PPP1R1B CTD-2330K ERCC8 FUT9 OTX2-AS1 CCDC101 STARD3 GNL3LP1 AL3564 9.3 CENPW NRXN3 SALL1 TCAP 58.1 MON1A NDUFAF2 MIR588 TRMT61A DDX19A PNMT SMIM15 MTF2 RBM6 RNU6-200P KLC1 CDH13 PGAP3 TMED5 RBM5 CTC-436P18.1 VIMP1 XRCC3 ERBB2 CTC-436P18.3 CCDC18 SEMA3F PRELID1P1 IGF2BP1 RNU1-1 USP34 CADM2 FBXL17 RPS4XP9 30P PSMD14 ALCAM UTRN MOV10 TBR1 SLC4A10 HNRNPA1P31 DPP4 FAT3 SPINK2 PWWP2AP1 SCAMP2 HIBADH PC2

Figure E19. Genetic correlation between educational attainment vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests1. The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for educational attainment are from a meta-analysis of genome- wide association data with N=293,723 individuals2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Okbay A, Beauchamp JP, Fontana MA, Lee JJ, Pers TH, Rietveld CA, Turley P, Chen GB, Emilsson V, Meddens SF, Oskarsson S. Genome-wide association study identifies 74 loci associated with educational attainment. Nature. 2016 May;533(7604):539.

A PC1 FOXN2 FAT3 ST7L PDE4D OPCML AKT3 BCL11A LINC01 SALL1 006

B CHD3 PC2

Figure E20. Genetic correlation between anorexia nervosa vs. surface area PC1 and PC2. Manhattan plots are shown for PC1 (panel A) and PC2 (panel B). The y-axis shows the 1 unadjusted –log10(P) values obtained from 1-sided rank-sum-based pleiotropy tests . The red dotted line indicates the genome-wide significance level of 5.0E-08. The genes mapped to the SNPs with P < 5.0E-08 are labelled and colored by chromosome numbers. The GWAS results for anorexia nervosa are from a meta-analysis of genome-wide association data with 2,907 cases with AN from 14 countries (15 sites) and 14,860 ancestrally matched controls as part of the Genetic Consortium for AN (GCAN) and the Wellcome Trust Case Control Consortium 3 (WTCCC3)2.

1Roshchupkin GV et al. Sum ranking: a new method to quantify pleiotropic loci for multiple phenotypes. (In preparation)

2Boraska V, Franklin CS, Floyd JA, Thornton LM, Huckins LM, Southam L, Rayner NW, Tachmazidou I, Klump KL, Treasure J, Lewis CM. A genome-wide association study of anorexia nervosa. Molecular Psychiatry. 2014 Oct;19(10):1085.