Iryna O. Fedko

The Genetics of Alcohol Dependence: Twin and SNP-Based Heritability, and Genome-Wide Association Study Based on AUDIT Scores

Alcohol dependence (AD) is among the most common and costly public health problems contributing to morbidity and mortality throughout the world. In this study, we investigate the genetic basis of AD in a Dutch population using data from the Netherlands Twin Register (NTR) and the Netherlands Study of Depression and Anxiety (NESDA). The presence of AD was ascertained via the Alcohol Use Disorders Identification Test (AUDIT) applying cut‐offs with good specificity and sensitivity in identifying those at risk for AD. Twin‐based heritability of AD‐AUDIT was estimated using structural equation modeling of data in 7,694 MZ and DZ twin pairs. Variance in AD‐AUDIT explained by all SNPs was estimated with genome‐wide complex trait analysis (GCTA). A genome‐wide association study (GWAS) was performed in 7,842 subjects. GWAS SNP effect concordance analysis was performed between our GWAS and a recent AD GWAS using DSM‐IV diagnosis. The twin‐based heritability of AD‐AUDIT was estimated at 60% (55–69%). GCTA showed that common SNPs jointly capture 33% (SE = 0.12, P = 0.002) of this heritability. In the GWAS, the top hits were positioned within four regions (4q31.1, 2p16.1, 6q25.1, 7p14.1) with the strongest association detected for rs55768019 (P = 7.58 × 10−7). This first GWAS of AD using the AUDIT measure found results consistent with previous genetic studies using DSM diagnosis: concordance in heritability estimates and direction of SNPs effect and overlap with top hits from previous GWAS. Thus, the use of appropriate questionnaires may represent cost‐effective strategies to phenotype samples in large‐scale biobanks or other population‐based datasets.

Short telomere length is associated with impaired cognitive performance in European ancestry cohorts

The association between telomere length (TL) dynamics on cognitive performance over the life-course is not well understood. This study meta-analyses observational and causal associations between TL and six cognitive traits, with stratifications on APOE genotype, in a Mendelian Randomization (MR) framework. Twelve European cohorts (N=17 052; mean age=59.2±8.8 years) provided results for associations between qPCR-measured TL (T/S-ratio scale) and general cognitive function, mini-mental state exam (MMSE), processing speed by digit symbol substitution test (DSST), visuospatial functioning, memory and executive functioning (STROOP). In addition, a genetic risk score (GRS) for TL including seven known genetic variants for TL was calculated, and used in associations with cognitive traits as outcomes in all cohorts. Observational analyses showed that longer telomeres were associated with better scores on DSST (β=0.051 per s.d.-increase of TL; 95% confidence interval (CI): 0.024, 0.077; P=0.0002), and MMSE (β=0.025; 95% CI: 0.002, 0.047; P=0.03), and faster STROOP (β=−0.053; 95% CI: −0.087, −0.018; P=0.003). Effects for DSST were stronger in APOE ɛ4 non-carriers (β=0.081; 95% CI: 0.045, 0.117; P=1.0 × 10−5), whereas carriers performed better in STROOP (β=−0.074; 95% CI: −0.140, −0.009; P=0.03). Causal associations were found for STROOP only (β=−0.598 per s.d.-increase of TL; 95% CI: −1.125, −0.072; P=0.026), with a larger effect in ɛ4-carriers (β=−0.699; 95% CI: −1.330, −0.069; P=0.03). Two-sample replication analyses using CHARGE summary statistics showed causal effects between TL and general cognitive function and DSST, but not with STROOP. In conclusion, we suggest causal effects from longer TL on better cognitive performance, where APOE ɛ4-carriers might be at differential risk.

Heritability and Genome-Wide Association Studies for Hair Color in a Dutch Twin Family Based Sample

Hair color is one of the most visible and heritable traits in humans. Here, we estimated heritability by structural equation modeling (N = 20,142), and performed a genome wide association (GWA) analysis (N = 7091) and a GCTA study (N = 3340) on hair color within a large cohort of twins, their parents and siblings from the Netherlands Twin Register (NTR). Self-reported hair color was analyzed as five binary phenotypes, namely “blond versus non-blond”, “red versus non-red”, “brown versus non-brown”, “black versus non-black”, and “light versus dark”. The broad-sense heritability of hair color was estimated between 73% and 99% and the genetic component included non-additive genetic variance. Assortative mating for hair color was significant, except for red and black hair color. From GCTA analyses, at most 24.6% of the additive genetic variance in hair color was explained by 1000G well-imputed SNPs. Genome-wide association analysis for each hair color showed that SNPs in the MC1R region were significantly associated with red, brown and black hair, and also with light versus dark hair color. Five other known genes (HERC2, TPCN2, SLC24A4, IRF4, and KITLG) gave genome-wide significant hits for blond, brown and light versus dark hair color. We did not find and replicate any new loci for hair color.

Psychopathology in 7‐year‐old children: Differences in maternal and paternal ratings and the genetic epidemiology

The assessment of childrens psychopathology is often based on parental report. Earlier studies have suggested that rater bias can affect the estimates of genetic, shared environmental and unique environmental influences on differences between children. The availability of a large dataset of maternal as well as paternal ratings of psychopathology in 7‐year old children enabled (i) the analysis of informant effects on these assessments, and (ii) to obtain more reliable estimates of the genetic and non‐genetic effects. DSM‐oriented measures of affective, anxiety, somatic, attention‐deficit/hyperactivity, oppositional‐defiant, conduct, and obsessive‐compulsive problems were rated for 12,310 twin pairs from the Netherlands Twin Register by mothers (N = 12,085) and fathers (N = 8,516). The effects of genetic and non‐genetic effects were estimated on the common and rater‐specific variance. For all scales, mean scores on maternal ratings exceeded paternal ratings. Parents largely agreed on the ranking of their childs problems (r 0.60–0.75). The heritability was estimated over 55% for maternal and paternal ratings for all scales, except for conduct problems (44–46%). Unbiased shared environmental influences, i.e., on the common variance, were significant for affective (13%), oppositional (13%), and conduct problems (37%). In clinical settings, different cutoffs for (sub)clinical scores could be applied to paternal and maternal ratings of their childs psychopathology. Only for conduct problems, shared environmental and genetic influences explain an equal amount in differences between children. For the other scales, genetic factors explain the majority of the variance, especially for the common part that is free of rater bias.

A method to customize population-specific arrays for genome-wide association testing

As an example of optimizing population-specific genotyping assays using a whole-genome sequence reference set, we detail the approach that followed to design the Axiom-NL array which is characterized by an improved imputation backbone based on the Genome of the Netherlands (GoNL) reference sequence and, compared with earlier arrays, a more comprehensive inclusion of SNPs on chromosomes X, Y, and the mitochondria. Common variants on the array were selected to be compatible with the Illumina Psych Array and the Affymetrix UK Biobank Axiom array. About 3.5% of the array (23 977 markers) represents SNPs from the GWAS catalog, including SNPs at FTO, APOE, Ion-channels, killer-cell immunoglobulin-like receptors, and HLA. Around 26 000 markers associated with common psychiatric disorders are included, as well as 6705 markers suggested to be associated with fertility and twinning. The platform can thus be used for risk profiling, detection of new variants, as well as ancestry determination. Results of coverage tests in 249 unrelated subjects with GoNL-based sequence data show that after imputation with 1000G as a reference, the median concordance between original and imputed genotypes is above 98%. The median imputation quality R2 for MAF thresholds of 0.001, 0.01, 0.05, and >0.05 are 0.05, 0.28, 0.80, 0.99, respectively, for the 1000G imputed SNPs, with a similar quality for the autosomes and X chromosome, showing a good genome-wide coverage for association studies after imputation.

Obsessive–compulsive symptoms in a large population-based twin-family sample are predicted by clinically based polygenic scores and by genome-wide SNPs

Variation in obsessive–compulsive symptoms (OCS) has a heritable basis, with genetic association studies starting to yield the first suggestive findings. We contribute to insights into the genetic basis of OCS by performing an extensive series of genetic analyses in a homogeneous, population-based sample from the Netherlands. First, phenotypic and genetic longitudinal correlations over a 6-year period were estimated by modeling OCS data from twins and siblings. Second, polygenic risk scores (PRS) for 6931 subjects with genotype and OCS data were calculated based on meta-analysis results from IOCDF-GC, to investigate their predictive value. Third, the contribution of measured single nucleotide polymorphisms (SNPs) to the heritability was estimated using random-effects modeling. Last, we performed an exploratory genome-wide association study (GWAS) of OCS, testing for SNP- and for gene-based associations. Stability in OCS (test–retest correlation 0.63) was mainly explained by genetic stability. The PRS based on clinical samples predicted OCS in our population-based twin-family sample. SNP-based heritability was estimated at 14%. GWAS revealed one SNP (rs8100480), located within the MEF2BNB gene, associated with OCS (P=2.56 × 10−8). Additional gene-based testing resulted in four significantly associated genes, which are located in the same chromosomal region on chromosome 19p13.11: MEF2BNB, RFXANK, MEF2BNB-MEF2B and MEF2B. Thus, common genetic variants explained a significant proportion of OCS trait variation. Genes significantly associated with OCS are expressed in the brain and involved in development and control of immune system functions (RFXANK) and regulation of gene expression of muscle-specific genes (MEF2BNB). MEF2BNB also showed a suggestive association with OCD in an independent case–control study, suggesting a role for this gene in the development of OCS.

Heritability and GWAS Studies for Monocyte-Lymphocyte Ratio

The monocyte-lymphocyte ratio (MLR) is a useful biomarker for disease development, but little is known about the extent to which genetic and environmental factors influence MLR variation. Here, we study the genetic architecture of MLR and determine the influence of demographic and lifestyle factors on MLR in data from a Dutch non-patient twin-family population. Data were obtained in 9,501 individuals from the Netherlands Twin Register. We used regression analyses to determine the effects of age, sex, smoking, and body mass index (BMI) on MLR and its subcomponents. Data on twins, siblings and parents (N = 7,513) were analyzed by genetic structural equation modeling to establish heritability and genome wide single nucleotide polymorphism (SNP) data from a genotyped subsample (N = 5,892) and used to estimate heritability explained by SNPs. SNP and phenotype data were also analyzed in a genome-wide association study to identify the genes involved in MLR. Linkage disequilibrium (LD) score regression and expression quantitative trait loci (eQTL) analyses were performed to further explore the genetic findings. Results showed that age, sex, and age × sex interaction effects were present for MLR and its subcomponents. Variation in MLR was not related to BMI, but smoking was positively associated with MLR. Heritability was estimated at 40% for MLR, 58% for monocyte, and 58% for lymphocyte count. The Genome-wide association study (GWAS) identified a locus on ITGA4 that was associated with MLR and only marginally significantly associated with monocyte count. For monocyte count, additional genetic variants were identified on ITPR3, LPAP1, and IRF8. For lymphocyte count, GWAS provided no significant findings. Taking all measured SNPs together, their effects accounted for 13% of the heritability of MLR, while all known and identified genetic loci explained 1.3% of variation in MLR. eQTL analyses showed that these genetic variants were unlikely to be eQTLs. In conclusion, variation in MLR level in the general population is heritable and influenced by age, sex, and smoking. We identified gene variants in the ITGA4 gene associated with variation in MLR. The significant SNP-heritability indicates that more genetic variants are likely to be involved.

Testing familial transmission of smoking with two different research designs

Introduction Classical twin studies show that smoking is heritable. To determine if shared family environment plays a role in addition to genetic factors, and if they interact (G×E), we use a children-of-twins design. In a second sample, we measure genetic influence with polygenic risk scores (PRS) and environmental influence with a question on exposure to smoking during childhood. Methods Data on smoking initiation were available for 723 children of 712 twins from the Netherlands Twin Register (64.9% female, median birth year 1985). Children were grouped in ascending order of risk, based on smoking status and zygosity of their twin-parent and his/her co-twin: never smoking twin-parent with a never smoking co-twin; never smoking twin-parent with a smoking dizygotic co-twin; never smoking twin-parent with a smoking monozygotic co-twin; and smoking twin-parent with a smoking or never smoking co-twin. For 4072 participants from the Netherlands Twin Register (67.3% female, median birth year 1973), PRS for smoking were computed and smoking initiation, smoking heaviness, and exposure to smoking during childhood were available. Results Patterns of smoking initiation in the four group children-of-twins design suggested shared familial influences in addition to genetic factors. PRS for ever smoking were associated with smoking initiation in all individuals. PRS for smoking heaviness were associated with smoking heaviness in individuals exposed to smoking during childhood, but not in non-exposed individuals. Conclusions Shared family environment influences smoking, over and above genetic factors. Genetic risk of smoking heaviness was only important for individuals exposed to smoking during childhood, versus those not exposed (G×E). Implications This study adds to the very few existing children-of-twins (CoT) studies on smoking and combines a CoT design with a second research design that utilizes polygenic risk scores and data on exposure to smoking during childhood. The results show that shared family environment affects smoking behavior over and above genetic factors. There was also evidence for gene-environment interaction (G×E) such that genetic risk of heavy versus light smoking was only important for individuals who were also exposed to (second-hand) smoking during childhood. Together, these findings give additional incentive to recommending parents not to expose their children to cigarette smoking.

Human pancreatic islet 3D chromatin architecture provides insights into the genetics of type 2 diabetes

Genetic studies promise to provide insight into the molecular mechanisms underlying type 2 diabetes (T2D). Variants associated with T2D are often located in tissue-specific enhancer regions (enhancer clusters, stretch enhancers or super-enhancers). So far, such domains have been defined through clustering of enhancers in linear genome maps rather than in 3D-space. Furthermore, their target genes are generally unknown. We have now created promoter capture Hi-C maps in human pancreatic islets. This linked diabetes-associated enhancers with their target genes, often located hundreds of kilobases away. It further revealed sets of islet enhancers, super-enhancers and active promoters that form 3D higher-order hubs, some of which show coordinated glucose-dependent activity. Hub genetic variants impact the heritability of insulin secretion, and help identify individuals in whom genetic variation of islet function is important for T2D. Human islet 3D chromatin architecture thus provides a framework for interpretation of T2D GWAS signals.

Predicting Loneliness with Polygenic Scores of Social, Psychological, and Psychiatric Traits

Loneliness is a heritable trait that accompanies multiple disorders. The association between loneliness and mental health indices may partly be due to inherited biological factors. We constructed polygenic scores for 27 traits related to behavior, cognition and mental health and tested their prediction for self‐reported loneliness in a population‐based sample of 8798 Dutch individuals. Polygenic scores for major depressive disorder (MDD), schizophrenia and bipolar disorder were significantly associated with loneliness. Of the Big Five personality dimensions, polygenic scores for neuroticism and conscientiousness also significantly predicted loneliness, as did the polygenic scores for subjective well‐being, tiredness and self‐rated health. When including all polygenic scores simultaneously into one model, only 2 major depression polygenic scores remained as significant predictors of loneliness. When controlling only for these 2 MDD polygenic scores, only neuroticism and schizophrenia remain significant. The total variation explained by all polygenic scores collectively was 1.7%. The association between the propensity to feel lonely and the susceptibility to psychiatric disorders thus pointed to a shared genetic etiology. The predictive power of polygenic scores will increase as the power of the genome‐wide association studies on which they are based increases and may lead to clinically useful polygenic scores that can inform on the genetic predisposition to loneliness and mental health.