Shared and Specific Associations of Amygdala Nuclei Volumes with PTSD Symptom Domains and Childhood Trauma: An ENIGMA-PGC PTSD Mega-Analysis (In Press, 2026)
Mufford et al., Imaging Workgroup
Study content & description coming soon.
Resting-state functional connectivity of the amygdala and hippocampus in PTSD – results from the PGC-ENIGMA PTSD working group (In Press, 2026)
Fig. 1. Network graph displaying significantly altered intrathalamic network edges among nuclei with significantly altered centrality in PTSD. Network edges with larger SCdiff appear redder and thicker (edge weights have been cubically scaled to emphasize differences). Nodes with a larger between-group strength difference appear redder/warmer in color.
Fig. 1: Deviations from Normative Subcortical Volume Associated with Childhood Trauma. A. Normative deviations in subcortical volumes associated with abuse in males and females; B. Normative deviations in subcortical volumes associated with neglect in males and females. β represents the standardized effect size from the general linear model (GLM, FDR corrected q<.05). Results from the pediatric cohort are not displayed as there were no significant associations between CTQ and subcortical volumes in this age cohort.
Fig. 1. Illustration of methods and results for static functional connectivity (SFC) and variability in FC (VFC). Methodology is on the left; results are on the right. Within-network (blue connecting lines) and between-network (green connecting lines) SFC and VFC were explored for the triple network of psychopathology for all listed diagnostic pairings. Brain figures have colored regions that are implicated in networks. Results of diagnosis × age interaction effects are illustrated on the right. There were significant differences in VFC between comorbid and control groups in salience network (SN)–default mode network (DMN) and SN–central executive network (CEN) connectivity. Within the SN, there were significant differences in VFC between comorbid participants and control participants and comorbid participants and participants with posttraumatic stress disorder (PTSD). FDR, false discovery rate; mTBI, mild traumatic brain injury.
Fig. 3. Coronal (left) and axial (right) views of thalamic RSFC between RSNs and the bilateral thalamus by contrasting cases > controls. Green boxes outline voxels that displayed significantly stronger RSFC between the default (top), salience (middle), and somatosensory (bottom) RSNs and thalamus in PTSD compared to control participants (pFWE < 0.05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 1: Schematic of IBMMA Workflow. Neuroimaging data from multiple subjects are flattened and segmented into memory-compatible subject-by-feature matrices. Statistical models are applied in parallel across features, and the resulting outputs are extracted and reconstructed to match the original data dimensions.
Fig. 1: Overview of the epigenome-wide associations with PTSD in immune cell types. Circular Manhattan plot indicating chromosomes on the outer track followed by the cell-type-specific associations, depicted as − log 10 (p-values), for B-cells (purple track), NK cells (blue track), CD8 + T cells (green track), and CD4 + T cells (yellow track). For each track, the red dashed line indicates the epigenome‐wide significance threshold of p < 9.0e-8. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 1. Early visual structural covariance is associated with posttraumatic stress disorder (PTSD) symptoms in the ENIGMA (Enhancing Neuro Imaging Genetics through Meta Analysis)-PGC (Psychiatric Genomics Consortium) cohort. Voxel-based morphometry (VBM) and pial surface area (PSA) features of a previously identified multimodal structural covariance network (SCN) (12) were projected onto data from the ENIGMA-PTSD and HCP-YA (Human Connectome Project–Young Adult) (A). Within the ENIGMA-PTSD dataset, individual participant loadings for the SCN were negatively correlated with PTSD symptoms (B); however, no relationship was observed between loadings and perceived stress in the HCP-YA dataset (C). Warmer and cooler colors in panel (A) reflect positive and negative relationship strength of each modality with spatial covariance loadings across regions, respectively. Graphs represent the partial plots from linear mixed effects models. Solid lines represent the unique association from the model, and the shaded bars represent the confidence intervals.
Fig. 1. Patients with PTSD exhibited lower regional gray matter volume compared to controls throughout the brain as seen in the orange highlighted regions in the figure, with a peak effect in the left cerebellum [−4,−72,-10] (see also Supplementary Table S6).
Fig. 1. Forest plot for the effect of the interaction term reflecting T1 Horvath DNAm age residuals by new-onset PTSD diagnosis on T2 Horvath DNAm age residuals (controlling for all other main effects in the model).
Fig. 1: Genetically Informed Brain Networks (GIBNs). Genomic structural equation modeling (gSEM) jointly modeled the genetic architecture of (A) cortical surface area (SA), and (B) cortical thickness, for 34 brain regions based on GWAS results of Grasby et al. [18]. The model generated 6 genetically informed brain networks (GIBNs) from SA phenotype measures. The color overlay on cortical regions represents the magnitude of the factor loadings indicated in the color gradient (yellow = high; blue = low). Subsequent GWAS identified several genome-wide significant hits (p < 5 × 10−8) associated with each GIBN.
Fig. 1. Genetic correlation and causal relationships linking posttraumatic stress disorder with other health outcomes. All phenotypes with a significant shared genetic causality (LCV GCP estimate significantly different from 0 after Bonferroni multiple testing correction) with PTSD are shown. A. PTSD genetic correlation with each phenotype. Red circles indicate genetic correlation, lines indicate 95% confidence intervals. B. Genetic causality proportion linking PTSD to each phenotype. Blue circles indicate the GCP shared between PTSD and a given phenotype, lines indicate the 95% CIs. The vertical dotted line indicates no shared causality. GCP estimates to the right of the dotted line indicate a causal influence of PTSD on the phenotype, values to the left the line indicate a causal influence of the phenotype on PTSD.
Fig. 1. The confusion matrix for Model 1 displays an accuracy of 92% on test data (N = 307), while the ROC curve indicates an AUC of 96% during the 10-fold cross-validation using all data (N = 1226).
Fig. 2: Summary of all analyses and findings. Figure combines CpGs from the main analysis (gold) and stratified analyses for sex, ancestry, and trauma type (light gold); and summarizes the results of blood and brain correlations (rose); cross-tissue associations for multiple brain regions (light blue), neuronal nuclei (blue), and a fibroblast model of prolonged stress (aqua), gene expression (purple); and genetic effects, including methylation quantitative trait loci (meQTL) analyses (light green) and genetic associations from the recent PGC-PTSD GWAS (dark green). Positive findings (p < 0.05) are indicated with the specific color of the respective category. Asterisk (*) indicates epigenome-wide significance (p < 9e-8). Gray represents the CpGs or genes were not present in the respective datasets. PFC: prefrontal cortex. EC: entorhinal cortex. STG: superior temporal gyrus. CER: cerebellum. dlPFC: dorsolateral prefrontal cortex. vmPFC: ventromedial prefrontal cortex. DG: dentate gyrus. 5mC: 5-Methylcytosine. 5hmC: 5-Hydroxymethylcytosine. GC: glucocorticoid.
Fig. 2: GWAS meta-analyses in European and multi-ancestry individuals identify a total of 95 PTSD risk loci. Overlaid Manhattan plots of EA (n = 137,136 cases and n = 1,085,746 controls) and multi-ancestry meta-analyses (n = 150,760 cases and n = 1,130,173 controls), showing 81 GWS loci for the EA (full circles) and 85 GWS loci for the multi-ancestry (hollow circles) analyses. Circle colors alternate between chromosomes, with even chromosomes colored blue and odd chromosomes colored black. The y axis refers to −log10(P) from two-sided z tests for meta-analysis effect estimates. The horizontal red bar indicates the threshold for GWS associations (P < 5 × 10−8).
Fig. 2: Genetic correlations between PTSD and immune-related phenotypes. Genetic correlations (rg) are indicated by circles that are drawn along the x axis. Phenotypes are colored by domain. Hollow circles indicate SNP based heritability (h2SNP) z-score <4 in the immune-related phenotype GWAS (rg estimates may be unreliable). The dotted vertical bar indicates the point of zero correlation. EGPA eosinophilic granulomatosis with polyangiitis.
Fig. 2:Effects of PTSD diagnosis on cerebellar subregion volumes. Atlas-based effect size (Cohen’s d) maps and MNI-based coronal slices (top: y = −72; bottom: y = −54) of the significant between-group differences for cerebellar subregion volumes in PTSD vs. Controls. Negative effect sizes reflect smaller volumes in PTSD. Regions significant at p-FDR < 0.05 are depicted in color, with the exception of right lobule V, where p-FDR = 0.051 after adjustment; right lobule V was significant p-FDR = 0.046) when examining PTSD severity instead of diagnosis. Grey-shaded subregions were non-significant. CM corpus medullary.
Fig. 1: Segmentation of the amygdala nuclei. (A) Using FreeSurfer (version 6.1), the amygdala was segmented into 9 nuclei: anterior amygdaloid area = yellow, cortico-amygdaloid transition area = dark blue, basal = red, lateral = light blue, accessory basal = orange, central = purple, medial = green. The cortical and paralaminar nuclei are not shown here. (B) Structural T1 scan provided for reference. Images provided by Morey et al. (12).
Fig. 3. Denoising Variational Autoencoder analysis pipeline: The model was trained using rs-fMRI or s-MRI data from controls only. The samples were then split into a training+validation (70 %) and independent-test (30 %) data. Then 20 % of the training data was set aside for validation and hyperparameter tuning. Once the training+validation was completed, the model’s performance was evaluated on the independent-test data, which provides an unbiased estimate of how the model generalizes to unseen data. The resulting VAE model learned to encode healthy patterns from the input brain features into its latent representation. Later, the brain features from patients with PTSD (PTSD test set) were input into the same VAE model, and the latent variables were extracted as new features for classification analysis.
Fig. 3. Regression analysis results revealed that PTSD diagnosis was associated with thinner cortex in multiple SCNs. The composite network visualization was obtained by assigning each vertex to the network that has the highest loading for that vertex (from the 𝑊 matrix), across all 20 networks. This association was maximal in SCNs 1 and 5, which included superior frontal cortex (SCN 1) and orbitofrontal cortex (SCN 5). Significant associations were also present in SCNs that included the motor cortex (SCN 2), insular cortex (SCN 4), medial superior frontal cortex (SCN 14), medial occipital cortex (SCN 15), anterior cingulate cortex (SCN 16), and posterior cingulate cortex (SCN 18). Both significant and non-significant SCNs are annotated with boundaries. Lateral and medial views of these significant SCNs are shown for left and right hemisphere, respectively. This figure is best viewed in color.
Fig. 1. Test area under the curve of (A) posttraumatic stress disorder vs. trauma exposed, (B) intrusive traumatic re-experiencing domain vs. trauma exposed; and (C) posttraumatic stress disorder vs. intrusive traumatic re-experiencing domain. FPR, false positive rate; TPR, true positive rate.
Fig. 3. The figure shows regions that showed increased centrality and decreased centrality in PTSD compared with controls for age groups: <10, 10–21, 22–39, 40–59, and >60 years.
Fig. 1.(A) Casual effect of attention deficit hyperactivity disorder (ADHD) on posttraumatic stress disorder (PTSD) phenotypes (left), and PTSD phenotypes on ADHD (right). All estimates lack evidence of heterogeneity and horizontal pleiotropy among the genetic instruments. For clarity, inverse variance weighted (IVW) and MR robust adjusted profile score (MR-RAPS, which accounts for weak genetic instruments) estimates are shown graphically and estimates across MR methods are provided in Table 4. Odds ratios and 95% confidence interval (CI) are reported for each MR test. The number of SNPs contributing to each genetic instrument is shown in parenthesis for each PTSD trait on the y-axis. Asterisks indicate significance after multiple testing correction accounting for all trait pairs and all MR causal inference methods applied (6 PTSD traits x 2 directions tested x 6 MR methods = 72 tests). (B) a directed acyclic diagram of multivariable MR results for ADHD and PTSD including bidirectional paths to covariates (green arrows) and direct causal paths to mediators (blue arrows) of the ADHD-PTSD relationship. All mediators and covariate traits were included in a single MVMR analysis. G indicates subsets of SNPs associated with one or more of the exposures. U indicates the effect of an unobserved confounder.
Study description coming soon.
Rare copy number variation in posttraumatic stress disorder (September 2022)
Maihofer et al., CNV Workgroup
Fig. 1: Genome-wide CNV burden association. The bar plot depicts regression beta coefficients as effect sizes (on the x-axis) of genome-wide CNV burden on PTSD, including overall burden, overlapping neurodevelopmental regions only, and genome-wide with neurodevelopmental regions excluded (on the y-axis). Data are shown stratified by CNV type, both CNV types (colored black), duplications only (colored red), and deletions only (colored blue). Effect sizes are shown in terms of megabases of the genome spanned by CNV.
Fig. 2. Site-specific cortical thickness averaged across regions. Non-harmonized (A), ComBat harmonized (B), and ComBat-GAM harmonized (C) data in participants with PTSD. Non-harmonized (D), ComBat harmonized (E), and ComBat-GAM harmonized (F) data in trauma-exposed controls. The order of sites in the figure is consistent with the order of site names in the legend from top to bottom to facilitate with interpretation. Compared to non-harmonized data, ComBat and ComBat-GAM lead to smaller differences between site-specific data and the mean values averaged across sites, and they do not change the site-specific standard deviations for cortical thickness. The effects of harmonization by LME models cannot be shown here because data harmonization and statistical analyses are inseparable in LME methods. Mean cortical thickness averaged across regions is shown to minimize regional biases. The boxplots were made using the default settings of the R ggplot2 function geom_boxplot(). The lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5 * IQR from the hinge (where IQR is the inter-quartile range, or distance between the first and third quartiles). The lower whisker extends from the hinge to the smallest value at most 1.5 * IQR of the hinge. Data beyond the end of the whiskers are called “outlying” points and are plotted individually.
Fig. 1. HALFpipeworkflow. HALFpipe is configured in a user interface where the user is asked a series of questions about their data and the processing steps to perform. Data are then converted to BIDS format (Gorgolewski et al., 2016) to allow standardized processing (white). After minimal preprocessing of the structural (blue) and functional (green and orange) data with fMRIPrep (Esteban, Blair, et al., 2019), additional preprocessing steps can be selected (red). Using the preprocessed data, statistical maps can be calculated during feature extraction (turquoise). Finally, group statistics can be performed (yellow). Note that not all preprocessing steps are available for each feature, as is outlined in Table 3. The diagram omits this information to increase visual clarity.
Fig. 2. The top-20 regions that (A) PTSD < non-PTSD and (B) PTSD > non-PTSD in cortical thickness. The top-20 regions that (C) PTSD < non-PTSD and (D) PTSD > non-PTSD in surface area. Node size represents the magnitude of effect size for between-group differences per region. Warm color denotes PTSD > non-PTSD, and cool color denotes PTSD < non-PTSD. Regions names are listed in Supplementary Table S4. Two examples are shown on the right to denote the node size and the corresponding effect size (Cohen’s d). The directions of the brain maps (axial view) are also shown.
Fig. 2: PTSS associates with DNAm across the genome.A) Manhattan plot for Meta-Analysis 1 across 3 cohorts (N samples = 858). Association analyses of each cohort are based on a random intercept model with a random effect of subject. B) Manhattan plot for Meta-Analysis 2 across 3 cohorts (N subjects = 429). Association analyses of each cohort are conducted by conditioning post-deployment methylation on baseline DNAm. The x-axis is the chromosomal location of each site across the genome. The y-axis is the −log10 of the unadjusted p-value for the association with PTSD symptom severity. The red line indicates genome-wide EWAS statistical significance at p < 9.0E-8.
Fig. 2. Comparison of the genetic correlations of posttraumatic stress disorder (PTSD) and lifetime trauma exposure (LTE) with other traits. The x-axis is the genetic correlation between LTE and a given trait from the LD Hub. The y-axis is the genetic correlation between PTSD and a given trait. Each dot depicts a given trait. Colored (black, red, or blue) dots indicate traits with significant genetic correlation to both PTSD and LTE after Bonferroni adjustment. Noncolored (gray) dots indicate traits where genetic correlation is not significant after Bonferroni adjustment. Blue dots indicate traits with significantly higher genetic correlation with PTSD than with LTE. Red dots indicate traits with significantly higher correlation with LTE than with PTSD. The top 5 traits with a significantly higher correlation to PTSD than LTE and top trait with significantly higher correlation to LTE than PTSD have been labeled.
Fig. 2. Latent causal variable (LCV) network among traumatic experiences, social support, and anthropometric traits surviving multiple testing correction. Panels A and B show male and female findings, respectively.
Fig. 1. Genetic correlation (rg) using Linkage Disequilibrium Score Regression (LDSC) (Bulik-Sullivan et al., 2015). Heatmap depicting rg between psychiatric disorders, where red denotes positive and blue denotes negative correlation estimates. The rows and columns of the heatmap are hierarchically clustered based on the correlation coefficients. The heritability (h2) estimates based on LDSC are given in the diagonal. Studies: 1) childhood trauma GWAS from Psychiatric Genomics Consortium (PGC) wave 2 (Ch.Trauma_PGCw2) (Dalvie et al., 2020); 2) major depression disorder PGC-GWAS wave 2 (MDD_PGCw2) (Wray et al., 2018), 3) GWAS of posttraumatic stress disorder (PTSD) in 21 European Civilian cohorts part of the PGC-PTSD wave 2 study (PTSD-CIV_PGCw2) (Nievergelt et al., 2019), 4) PTSD-GWAS in 20 European military cohorts part of the PGC-PTSD wave 2 study (PTSD-MIL_PGCw2), and 5) anxiety GWAS from PGC wave 1 (Anxiety_PGCw1) (Otowa et al., 2016a). For comparison, two additional studies are included: schizophrenia (SCZ_PGCw2) (Schizophrenia Working Group of the Psychiatric Genomics et al., 2014) and bipolar disorder (BIP_PGCw2) (Stahl et al., 2019) based on PGC wave 2. Note that h2 is not reported on a liability scale. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Fig. 2. Co-expression modules detected in NHW MIRECC/Duke subjects and subsequent associations with PTSD, PTSD severity, MDD, and smoking status. The heatmap color scale represents effect sizes of each association (betas are listed in each cell, followed by corresponding p-value in parenthesis).
Fig. 3. Forest plots of trans-ethnic meta-analyses for the main effect of blood pressure polygenic scores on (A) systolic blood pressure and (B) diastolic blood pressure. For each cohort, a square is plotted at the effect estimate value. The size of each plotted square reflects relative precision, where size is inversely proportional to the standard error of a given cohort. Reported effects are beta coefficients and 95% confidence intervals. AA = African ancestry; EA = European ancestry; FE = fixed effects.
Fig. 1. Light blue indicates regions with smaller volume in PTSD group. Dark blue indicates regions which are smaller in PTSD group, and their volumes are negatively associated with harmonized PTSS severity scores.
Fig. 1. Manhattan plot for the meta-analysis of 10 cohorts (Ncases = 878, Ncontrols = 1,018). Association analyses of each cohort are based on empirical Bayes method. Meta-analysis is done by sample size-weighted sum of z-scores. p-values are two-sided and unadjusted for multiple testing. All of the association models are adjusted for sex (if applicable), age, CD8, CD4, NK, B cell, and monocyte cell proportions, and ancestry using principal components (PCs). The x-axis is the location of each site across the genome. The y-axis is the –log10 of the p-value for the association with PTSD. The dashed line indicates statistical significance at p < 3.6E-8.
Fig. 1. (A–C) Manhattan plots showing the −log10 of the p value of association of the genes-by-tissue in each chomosome with PTSD. In (A), one association reached study-wide significance (depicted by red discontinuous line) in our overall EA meta-analysis (23,195 cases/151,447 controls), and two genes reached within-tissue significance (depicted by purple discontinuous line). In (B), one gene reached study-wide significance in the military-only EA meta-analysis (6,004 cases/21,534 controls), and one reached within-tissue significance, while (C) two genes reached within-tissue significance in the civilian-only EA meta-analysis (16,959 cases/129,607 controls). Gene-tissue pairs with p ⩽ 10-4 are color coded according to tissue type.
Table 1. Heritability estimates from LDSC regression analyses. Note:* = Liability scales for PTSD used a sample prevalence of 28% and a population prevalence of 30%, as done in previously published work with this dataset. Liability scales for AD used a sample prevalence of 35% and a population prevalence of 16% as done in previously published work with this dataset.
Fig. 2: Genetic correlation (x-axis: genetic correlation, rg; y-axis: –log10p-value) of serum C-reactive protein (CRP) with post-traumatic stress disorder (PTSD), trauma response, trauma exposure, social support, and socioeconomic status. Full circles represent nominally significant genetic correlations (p < 0.05). As expected, CRP showed positive genetic correlation with PTSD, trauma response, trauma exposure and deprivation index, and negative genetic correlation with social support and household income.
Fig. 2. Box/Scatter plot of the proportion of DNAm (Beta) for (a) cg19534438, the genome-wide significant locus in G0S2, (b) cg05575921, the smoking- and PTSD-associated locus in AHRR, and (c) the CHST11 peak locus from the EWAS of whole-blood samples that was corrected significant in the analysis of tissue from the PFC, with blood, dlPFC, and vmPFC methylation plotted separately.
Fig. 1. Manhattan plot of UKBB GWAS for childhood maltreatment, showing the top variants. The horizontal line represents genome-wide significance at p < 5 × 10−8.
Fig. 1. Manhattan plot showing the results of the stage 2 meta-analysis across 3 epigenome-wide association studies (MRS, Army STARRS, PRISMO). The upper part shows the 3 significant differentially methylated positions (DMPs) while the lower part shows the 12 significant differentially methylated regions (DMRs). Red lines indicate significance thresholds after the Bonferroni corrections for ~ 485,000 (top) and 26,000 (bottom) comparisons, respectively.
Fig. 1. Genetic correlations (rg) between sleep phenotypes and posttraumatic stress disorder (PTSD). Genetic correlation estimates obtained through linkage disequilibrium score regression (LDSC) are shown here, grouped by sleep phenotype, with each specific study name on the left. Error bars represent standard error. p-Values that pass multiple testing correction (p < 0.0022 per Bonferroni correction) are indicated by an (*).
Fig. 3. Tapetum displayed on the ENIGMA template FA. The skeleton is shown in red, the left tapetum (green) and right tapetum (blue) ROIs are displayed. Left in image is right in brain.
Fig. 1. Manhattan plots from meta-analyses of PTSD GWAS, showing the top variants in six independent genome-wide significant loci. Results are shown for subjects of European (EUA; a) and African ancestry (AFA; c), and for sex-stratified analyses in EUA men (b) and AFA men (d), respectively. Sex-stratified analyses for women were not significant (Supplementary Fig. 4). The red line represents genome-wide significance at P < 5 × 10−8. Note: rs148757321 and rs142174523 do not remain significant after Bonferroni-adjustment for sex-stratified analyses (at P < 1.67 × 10−8)
Polimanti et al., GWAS & Systems Biology Workgroups
Fig. 1: Single-Nucleotide Polymorphism (SNP) Repeated Effects on Posttraummatic Stress Disorder (PTSD) and Most Advanced Math Course Completed (MC). SNP exposure (MC associations, β) and SNP outcome (PTSD freeze-2 associations, log odds ratio [OR]) coefficients used in the mendelian randomization analysis. Crosses represent 95% CIs for each association.
Fig. 2. Forest plot showing the β-coefficients for the effect of current post-traumatic stress disorder on methylation at cg23637605 (NRG1) from the linear regression within each cohort and the combined effect from the meta-analysis in our primary analysis.
Fig. 2.Main interactions and dataflow among the nine Psychiatric Genomics Consortium posttraumatic stress disorder (PTSD) working groups. The Psychiatric Genomics Consortium PTSD genome-wide association study (GWAS) group has drastically expanded its scope since its initiation in 2013. It currently includes working groups with emphases on complementary phenotypes (psychophysiology, physical health, and imaging), working groups contributing complementary “omics” data (copy number variants [CNV], epigenetics, transcriptome, and microbiome), and a systems biology group aiming at integration of the different types of data. Arrows indicate primary flow of data, but interactions among groups are expanding. EWAS, epigenome-wide association study; MRI, magnetic resonance imaging; RNAseq, RNA sequencing; SNP, single nucleotide polymorphism.
Fig. 3. Shows the forest plots for the significant meta-analytic associations between childhood trauma (top panel) and lifetime PTSD severity (bottom panel) and Hannum DNAm age residuals. Error bars represent 95% confidence intervals. Study abbreviations are defined in Table 1.
Fig. 1. SNP-exposure (WCadj associations) and SNP-outcome (PTSD associations) coefficients used in the MR analysis. Error bars (95% CIs) are reported for each association. The solid line represents the inverse-variance-weighted estimate
Fig. 3. Ancestry inference using SNPs versus methylation probes in 128 participants of the Marine Resiliency Study (MRS). (A) Principal component (PC) plot showing ancestry inferred using SNPs from a genome-wide association study (GWAS). PC plots based on CpG probes with SNPs within 1 bp distance (B) and with SNPs within 10 bp distance (C), respectively. Subject are placed into four ancestral groups based on ancestry estimates using ancestry-informative SNPs and a reference panel (see methods).
Fig. 3. PTSD SNP-chip heritability (h2SNP) overall and for males and females separately and comparison with other psychiatric disorders. Gray bars denote PTSD heritability estimates. Slashed bars reflect SCZ, BIP and MDD heritability estimates calculated using LDSC as applied to published data.16, 31, 32 Red lines denote twin study heritability estimates, see Discussion. European-American (EA) samples only per description in text; error bars reflect s.e. BIP, bipolar disorder; LDSC, linkage disequilibrium score regression; MDD, major depressive disorder; PTSD, posttraumatic stress disorder; SCZ, schizophrenia; SNP, single-nucleotide polymorphism.
Fig. 1. A comparison of ancestral diversity in (a) representative Psychiatric Genomics Consortium (PGC) samples of primarily European ancestry and (b) representative PGC–PTSD samples. Key: mrsa, mrsb—subsets (a and b) of the Marine Resilience Study (Nievergelt); gtpx—Grady Trauma Project (Ressler); gsdx—Genetics of Substance Dependence (Gelernter); fscd—Family Studies of Cocaine Dependence (Bierut); dnhs—Detroit Neighborhood Health Study (Aiello); cogb, coga—subsets (a and b) of the COGEND study (Bierut); Note that African American refers to subjects from the USA who typically have a mix of African and European ancestry, whereas African Ancestry refers to subjects from Africa without admixed ancestry. PTSD, posttraumatic stress disorder.
