Integration of risk factor polygenic risk score with disease polygenic risk score for disease prediction.

Polygenic risk score (PRS) is useful for capturing an individual's genetic susceptibility. However, previous studies have not fully exploited the potential of the risk factor PRS (RFPRS) for disease prediction. We explored the potential of integrating disease-related RFPRSs with disease PRS to enhance disease prediction performance. We constructed 112 RFPRSs and analyzed the association of RFPRSs with diseases to identify disease-related RFPRSs in 700 diseases, using the UK Biobank dataset. We uncovered 6157 statistically significant associations between 247 diseases and 109 RFPRSs. We estimated the disease PRSs of 70 diseases that exhibited statistically significant heritability, to generate RFDiseasemetaPRS-a combined PRS integrating RFPRSs and disease PRS-and compare the prediction performance metrics between RFDiseasemetaPRS and disease PRS. RFDiseasemetaPRS showed better performance for Nagelkerke's pseudo-R2, odds ratio (OR) per 1 SD, net reclassification improvement (NRI) values and difference of R2 considered by variance of R2 in 31 out of 70 diseases. Additionally, we assessed risk classification between two models by examining OR between the top 10% and remaining 90% individuals for the 31 diseases; RFDiseasemetaPRS exhibited better R2, NRI and OR than disease PRS. These findings highlight the importance of utilizing RFDiseasemetaPRS, which can provide personalized healthcare and tailored prevention strategies.

Identification of asthma-related genes using asthmatic blood eQTLs of Korean patients.

BACKGROUND: More than 200 asthma-associated genetic variants have been identified in genome-wide association studies (GWASs). Expression quantitative trait loci (eQTL) data resources can help identify causal genes of the GWAS signals, but it can be difficult to find an eQTL that reflects the disease state because most eQTL data are obtained from normal healthy subjects. METHODS: We performed a blood eQTL analysis using transcriptomic and genotypic data from 433 Korean asthma patients. To identify asthma-related genes, we carried out colocalization, Summary-based Mendelian Randomization (SMR) analysis, and Transcriptome-Wide Association Study (TWAS) using the results of asthma GWASs and eQTL data. In addition, we compared the results of disease eQTL data and asthma-related genes with two normal blood eQTL data from Genotype-Tissue Expression (GTEx) project and a Japanese study. RESULTS: We identified 340,274 cis-eQTL and 2,875 eGenes from asthmatic eQTL analysis. We compared the disease eQTL results with GTEx and a Japanese study and found that 64.1% of the 2,875 eGenes overlapped with the GTEx eGenes and 39.0% with the Japanese eGenes. Following the integrated analysis of the asthmatic eQTL data with asthma GWASs, using colocalization and SMR methods, we identified 15 asthma-related genes specific to the Korean asthmatic eQTL data. CONCLUSIONS: We provided Korean asthmatic cis-eQTL data and identified asthma-related genes by integrating them with GWAS data. In addition, we suggested these asthma-related genes as therapeutic targets for asthma. We envisage that our findings will contribute to understanding the etiological mechanisms of asthma and provide novel therapeutic targets.

Genetic differences according to onset age and lung function in asthma: A cluster analysis.

BACKGROUND: The extent of differences between genetic risks associated with various asthma subtypes is still unknown. To better understand the heterogeneity of asthma, we employed an unsupervised method to identify genetic variants specifically associated with asthma subtypes. Our goal was to gain insight into the genetic basis of asthma. METHODS: In this study, we utilized the UK Biobank dataset to select asthma patients (All asthma, n = 50,517) and controls (n = 283,410). We excluded 14,431 individuals who had no information on predicted values of forced expiratory volume in one second percent (FEV1%) and onset age, resulting in a final total of 36,086 asthma cases. We conducted k-means clustering based on asthma onset age and predicted FEV1% using these samples (n = 36,086). Cluster-specific genome-wide association studies were then performed, and heritability was estimated via linkage disequilibrium score regression. To further investigate the pathophysiology, we conducted eQTL analysis with GTEx and gene-set enrichment analysis with FUMA. RESULTS: Clustering resulted in four distinct clusters: early onset asthmanormalLF (early onset with normal lung function, n = 8172), early onset asthmareducedLF (early onset with reduced lung function, n = 8925), late-onset asthmanormalLF (late-onset with normal lung function, n = 12,481), and late-onset asthmareducedLF (late-onset with reduced lung function, n = 6508). Our GWASs in four clusters and in All asthma sample identified 5 novel loci, 14 novel signals, and 51 cluster-specific signals. Among clusters, early onset asthmanormalLF and late-onset asthmareducedLF were the least correlated (rg  = 0.37). Early onset asthmareducedLF showed the highest heritability explained by common variants (h2  = 0.212) and was associated with the largest number of variants (71 single nucleotide polymorphisms). Further, the pathway analysis conducted through eQTL and gene-set enrichment analysis showed that the worsening of symptoms in early onset asthma correlated with lymphocyte activation, pathogen recognition, cytokine receptor activation, and lymphocyte differentiation. CONCLUSIONS: Our findings suggest that early onset asthmareducedLF was the most genetically predisposed cluster, and that asthma clusters with reduced lung function were genetically distinct from clusters with normal lung function. Our study revealed the genetic variation between clusters that were segmented based on onset age and lung function, providing an important clue for the genetic mechanism of asthma heterogeneity.

Investigation of heteroscedasticity in polygenic risk scores across 15 quantitative traits.

The polygenic risk score (PRS) could be used to stratify individuals with high risk of diseases and predict complex trait of individual in a population. Previous studies developed a PRS-based prediction model using linear regression and evaluated the predictive performance of the model using the R 2 value. One of the key assumptions of linear regression is that the variance of the residual should be constant at each level of the predictor variables, called homoscedasticity. However, some studies show that PRS models exhibit heteroscedasticity between PRS and traits. This study analyzes whether heteroscedasticity exists in PRS models of diverse disease-related traits and, if any, it affects the accuracy of PRS-based prediction in 354,761 Europeans from the UK Biobank. We constructed PRSs for 15 quantitative traits using LDpred2 and estimated the existence of heteroscedasticity between PRSs and 15 traits using three different tests of the Breusch-Pagan (BP) test, score test, and F test. Thirteen out of fifteen traits show significant heteroscedasticity. Further replication using new PRSs from the PGS catalog and independent samples (N = 23,620) from the UK Biobank confirmed the heteroscedasticity in ten traits. As a result, ten out of fifteen quantitative traits show statistically significant heteroscedasticity between the PRS and each trait. There was a greater variance of residuals as PRS increased, and the prediction accuracy at each level of PRS tended to decrease as the variance of residuals increased. In conclusion, heteroscedasticity was frequently observed in the PRS-based prediction models of quantitative traits, and the accuracy of the predictive model may differ according to PRS values. Therefore, prediction models using the PRS should be constructed by considering heteroscedasticity.

Gene-environment interaction explains a part of missing heritability in human body mass index.

Gene-environment (G×E) interaction could partially explain missing heritability in traits; however, the magnitudes of G×E interaction effects remain unclear. Here, we estimate the heritability of G×E interaction for body mass index (BMI) by subjecting genome-wide interaction study data of 331,282 participants in the UK Biobank to linkage disequilibrium score regression (LDSC) and linkage disequilibrium adjusted kinships-software for estimating SNP heritability from summary statistics (LDAK-SumHer) analyses. Among 14 obesity-related lifestyle factors, MET score, pack years of smoking, and alcohol intake frequency significantly interact with genetic factors in both analyses, accounting for the partial variance of BMI. The G×E interaction heritability (%) and standard error of these factors by LDSC and LDAK-SumHer are as follows: MET score, 0.45% (0.12) and 0.65% (0.24); pack years of smoking, 0.52% (0.13) and 0.93% (0.26); and alcohol intake frequency, 0.32% (0.10) and 0.80% (0.17), respectively. Moreover, these three factors are partially validated for their interactions with genetic factors in other obesity-related traits, including waist circumference, hip circumference, waist-to-hip ratio adjusted with BMI, and body fat percentage. Our results suggest that G×E interaction may partly explain the missing heritability in BMI, and two G×E interaction loci identified could help in understanding the genetic architecture of obesity.

The effect of heteroscedasticity on the prediction efficiency of genome-wide polygenic score for body mass index.

Globally, more than 1.9 billion adults are overweight. Thus, obesity is a serious public health issue. Moreover, obesity is a major risk factor for diabetes mellitus, coronary heart disease, and cardiovascular disease. Recently, GWAS examining obesity and body mass index (BMI) have increasingly unveiled many aspects of the genetic architecture of obesity and BMI. Information on genome-wide genetic variants has been used to estimate the genome-wide polygenic score (GPS) for a personalized prediction of obesity. However, the prediction power of GPS is affected by various factors, including the unequal variance in the distribution of a phenotype, known as heteroscedasticity. Here, we calculated a GPS for BMI using LDpred2, which was based on the BMI GWAS summary statistics from a European meta-analysis. Then, we tested the GPS in 354,761 European samples from the UK Biobank and found an effective prediction power of the GPS on BMI. To study a change in the variance of BMI, we investigated the heteroscedasticity of BMI across the GPS via graphical and statistical methods. We also studied the homoscedastic samples for BMI compared to the heteroscedastic sample, randomly selecting samples with various standard deviations of BMI residuals. Further, we examined the effect of the genetic interaction of GPS with environment (GPS×E) on the heteroscedasticity of BMI. We observed the changing variance (i.e., heteroscedasticity) of BMI along the GPS. The heteroscedasticity of BMI was confirmed by both the Breusch-Pagan test and the Score test. Compared to the heteroscedastic sample, the homoscedastic samples from small standard deviation of BMI residuals showed a decreased heteroscedasticity and an improved prediction accuracy, suggesting a quantitatively negative correlation between the phenotypic heteroscedasticity and the prediction accuracy of GPS. To further test the effects of the GPS×E on heteroscedasticity, first we tested the genetic interactions of the GPS with 21 environments and found 8 significant GPS×E interactions on BMI. However, the heteroscedasticity of BMI was not ameliorated after adjusting for the GPS×E interactions. Taken together, our findings suggest that the heteroscedasticity of BMI exists along the GPS and is not affected by the GPS×E interaction.

Identification of five genetic variants with differential effects on obesity-related traits based on age.

Obesity is a major public health concern, and its prevalence generally increases with age. As the number of elderly people is increasing in the aging population, the age-dependent increase in obesity has raised interest in the underlying mechanism. To understand the genetic basis of age-related increase in obesity, we identified genetic variants showing age-dependent differential effects on obesity. We conducted stratified analyses between young and old groups using genome-wide association studies of 355,335 United Kingom Biobank participants for five obesity-related phenotypes, including body mass index, body fat percentage, waist-hip ratio, waist circumference, and hip circumference. Using t-statistic, we identified five significant lead single nucleotide polymorphisms: rs2258461 with body mass index, rs9861311 and rs429358 with body fat percentage, rs2870099 with waist-hip ratio, and rs145500243 with waist circumference. Among these single nucleotide polymorphisms, rs429358, located in APOE gene was associated with diverse age-related diseases, such as Alzheimer's disease, coronary artery disease, age-related degenerative macular diseases, and cognitive decline. The C allele of rs429358 gradually decreases body fat percentage as one grows older in the range of 40-69 years. In conclusion, we identified five genetic variants with differential effects on obesity-related phenotypes based on age using a stratified analysis between young and old groups, which may help to elucidate the mechanisms by which age influences the development of obesity.

Smoking-Interaction Loci Affect Obesity Traits: A Gene-Smoking Stratified Meta-Analysis of 545,131 Europeans.

INTRODUCTION: Although many studies have investigated the association between smoking and obesity, very few have analyzed how obesity traits are affected by interactions between genetic factors and smoking. Here, we aimed to identify the loci that affect obesity traits via smoking status-related interactions in European samples. METHODS: We performed stratified analysis based on the smoking status using both the UK Biobank (UKB) data (N = 334,808) and the Genetic Investigation of ANthropometric Traits (GIANT) data (N = 210,323) to identify gene-smoking interaction for obesity traits. We divided the UKB subjects into two groups, current smokers and nonsmokers, based on the smoking status, and performed genome-wide association study (GWAS) for body mass index (BMI), waist circumference adjusted for BMI (WCadjBMI), and waist-hip ratio adjusted for BMI (WHRadjBMI) in each group. And then we carried out the meta-analysis using both GWAS summary statistics of UKB and GIANT for BMI, WCadjBMI, and WHRadjBMI and computed the stratified p values (pstratified) based on the differences between meta-analyzed estimated beta coefficients with standard errors in each group. RESULTS: We identified four genome-wide significant loci in interactions with the smoking status (pstratified < 5 × 10-8): rs336396 (INPP4B) and rs12899135 (near CHRNB4) for BMI, and rs998584 (near VEGFA) and rs6916318 (near RSPO3) for WHRadjBMI. Moreover, we annotated the biological functions of the SNPs using expression quantitative trait loci (eQTL) and GWAS databases, along with publications, which revealed possible mechanisms underlying the association between the smoking status-related genetic variants and obesity. CONCLUSIONS: Our findings suggest that obesity traits can be modified by the smoking status via interactions with genetic variants through various biological pathways.

A cardiac-null mutation of Prdm16 causes hypotension in mice with cardiac hypertrophy via increased nitric oxide synthase 1.

Hypertension or hypotension prevails as a comorbidity in patients with heart failure (HF). Although blood pressure (BP) is an important factor in managing the mortality of HF, the molecular mechanisms of changes in BP have not been clearly understood in cases of HF. We and others have demonstrated that a loss in PRDM16 causes hypertrophic cardiomyopathy, leading to HF. We aimed to determine whether BP is altered in mice that experience cardiac loss of Prdm16 and identify the underlying mechanism of BP-associated changes. BP decreased significantly only in female mice with a cardiac-null mutation of Prdm16 compared with controls, by an invasive protocol under anesthesia and by telemetric method during conscious, unrestrained status. Mice with a cardiac loss of Prdm16 had higher heart-to-body weight ratios and upregulated atrial natriuretic peptide, suggesting cardiac hypertrophy. Plasma aldosterone-to-renin activity ratios and plasma sodium levels decreased in Prdm16-deficient mice versus control. By RNA-seq and in subsequent functional analyses, Prdm16-null hearts were enriched in factors that regulate BP, including Adra1a, Nos1, Nppa, and Nppb. The inhibition of nitric oxide synthase 1 (NOS1) reverted the decrease in BP in cardiac-specific Prdm16 knockout mice. Mice with cardiac deficiency of Prdm16 present with hypotension and cardiac hypertrophy. Further, our findings suggest that the increased expression of NOS1 causes hypotension in mice with a cardiac-null mutation of Prdm16. These results provide novel insights into the molecular mechanisms of hypotension in subjects with HF and contribute to our understanding of how hypotension develops in patients with HF.

Genome-Wide Interaction Study of Late-Onset Asthma With Seven Environmental Factors Using a Structured Linear Mixed Model in Europeans.

Asthma is among the most common chronic diseases worldwide, creating a substantial healthcare burden. In late-onset asthma, there are wide global differences in asthma prevalence and low genetic heritability. It has been suggested as evidence for genetic susceptibility to asthma triggered by exposure to multiple environmental factors. Very few genome-wide interaction studies have identified gene-environment (G×E) interaction loci for asthma in adults. We evaluated genetic loci for late-onset asthma showing G×E interactions with multiple environmental factors, including alcohol intake, body mass index, insomnia, physical activity, mental status, sedentary behavior, and socioeconomic status. In gene-by-single environment interactions, we found no genome-wide significant single-nucleotide polymorphisms. However, in the gene-by-multi-environment interaction study, we identified three novel and genome-wide significant single-nucleotide polymorphisms: rs117996675, rs345749, and rs17704680. Bayes factor analysis suggested that for rs117996675 and rs17704680, body mass index is the most relevant environmental factor; for rs345749, insomnia and alcohol intake frequency are the most relevant factors in the G×E interactions of late-onset asthma. Functional annotations implicate the role of these three novel loci in regulating the immune system. In addition, the annotation for rs117996675 supports the body mass index as the most relevant environmental factor, as evidenced by the Bayes factor value. Our findings help to understand the role of the immune system in asthma and the role of environmental factors in late-onset asthma through G×E interactions. Ultimately, the enhanced understanding of asthma would contribute to better precision treatment depending on personal genetic and environmental information.

Gene-environment interaction in type 2 diabetes in Korean cohorts: Interaction of a type 2 diabetes polygenic risk score with triglyceride and cholesterol on fasting glucose levels.

Type 2 diabetes (T2D) is caused by genetic and environmental factors as well as gene-environment interactions. However, these interactions have not been systematically investigated. We analyzed these interactions for T2D and fasting glucose levels in three Korean cohorts, HEXA, KARE, and CAVAS, using the baseline data with a multiple regression model. Two polygenic risk scores for T2D (PRST2D ) and fasting glucose (PRSFG ) were calculated using 488 and 82 single nucleotide polymorphisms (SNP) for T2D and fasting glucose, respectively, which were extracted from large-scaled genome-wide association studies with multiethnic data. Both lifestyle risk factors and T2D-related biochemical measurements were assessed. The effect of interactions between PRST2D -triglyceride (TG) and PRST2D -total cholesterol (TC) on fasting glucose levels was observed as follows: ß ± SE = 0.0005 ± 0.0001, p = 1.06 × 10-19 in HEXA, ß ± SE = 0.0008 ± 0.0001, p = 2.08 × 10-8 in KARE for TG; ß ± SE = 0.0006 ± 0.0001, p = 2.00 × 10-6 in HEXA, ß ± SE = 0.0020 ± 0.0004, p = 2.11 × 10-6 in KARE, ß ± SE = 0.0007 ± 0.0004, p = 0.045 in CAVAS for TC. PRST2D -based classification of the participants into four groups showed that the fasting glucose levels in groups with higher PRST2D were more adversely affected by both the TG and TC. In conclusion, blood TG and TC levels may affect the fasting glucose level through interaction with T2D genetic factors, suggesting the importance of consideration of gene-environment interaction in the preventive medicine of T2D.

Characterisation of insomnia as an environmental risk factor for asthma via Mendelian randomization and gene environment interaction.

Asthma is a complex disease that is reportedly associated with insomnia. However, the causal directionality of this association is still unclear. We used asthma and insomnia-associated single nucleotide polymorphisms (SNPs) and genome-wide association study (GWAS) summary statistics to test the causal directionality between insomnia and asthma via Mendelian randomization (MR) analysis. We also performed a cross-trait meta-analysis using UK Biobank GWAS summary statistics and a gene-environment interaction study using data from UK Biobank. The interaction of genetic risk score for asthma (GRSasthma) with insomnia on asthma was tested by logistic regression. Insomnia was a risk factor for the incidence of asthma, as revealed by three different methods of MR analysis. However, asthma did not act as a risk factor for insomnia. The cross-trait meta-analysis identified 28 genetic loci shared between asthma and insomnia. In the gene-environment interaction study, GRSasthma interacted with insomnia to significantly affect the risk of asthma. The results of this study highlight the importance of insomnia as a risk factor of asthma, and warrant further analysis of the mechanism through which insomnia affects the risk of asthma.

Association Between Environmental Factors and Asthma Using Mendelian Randomization: Increased Effect of Body Mass Index on Adult-Onset Moderate-to-Severe Asthma Subtypes.

Although asthma is one of the most common chronic diseases throughout all age groups, its etiology remains unknown, primarily due to its heterogeneous characteristics. We examined the causal effects of various environmental factors on asthma using Mendelian randomization and determined whether the susceptibility to asthma due to the causal effect of a risk factor differs between asthma subtypes, based on age of onset, severity of asthma, and sex. We performed Mendelian randomization analyses (inverse variance weighted, weighted median, and generalized summary-data-based Mendelian randomization) using UK Biobank data to estimate the causal effects of 69 environmental factors on asthma. Additional sensitivity analyses (MR-Egger regression, Cochran's Q test, clumping, and reverse Mendelian randomization) were performed to ensure minimal or no pleiotropy. For confirmation, two-sample setting analyses were replicated using BMI SNPs that had been reported by a meta-genome-wide association study in Japanese and European (GIANT) populations and a genome-wide association study in control individuals from the UK Biobank. We found that BMI causally affects the development of asthma and that the adult-onset moderate-to-severe asthma subtype is the most susceptible to causal inference by BMI. Further, it is likely that the female subtype is more susceptible to BMI than males among adult asthma cases. Our findings provide evidence that obesity is a considerable risk factor in asthma patients, particularly in adult-onset moderate-to-severe asthma cases, and that weight loss is beneficial for reducing the burden of asthma.

Identification of genetic loci affecting body mass index through interaction with multiple environmental factors using structured linear mixed model.

Multiple environmental factors could interact with a single genetic factor to affect disease phenotypes. We used Struct-LMM to identify genetic variants that interacted with environmental factors related to body mass index (BMI) using data from the Korea Association Resource. The following factors were investigated: alcohol consumption, education, physical activity metabolic equivalent of task (PAMET), income, total calorie intake, protein intake, carbohydrate intake, and smoking status. Initial analysis identified 7 potential single nucleotide polymorphisms (SNPs) that interacted with the environmental factors (P value < 5.00 × 10-6). Of the 8 environmental factors, PAMET score was excluded for further analysis since it had an average Bayes Factor (BF) value < 1 (BF = 0.88). Interaction analysis using 7 environmental factors identified 11 SNPs (P value < 5.00 × 10-6). Of these, rs2391331 had the most significant interaction (P value = 7.27 × 10-9) and was located within the intron of EFNB2 (Chr 13). In addition, the gene-based genome-wide association study verified EFNB2 gene significantly interacting with 7 environmental factors (P value = 5.03 × 10-10). BF analysis indicated that most environmental factors, except carbohydrate intake, contributed to the interaction of rs2391331 on BMI. Although the replication of the results in other cohorts is warranted, these findings proved the usefulness of Struct-LMM to identify the gene-environment interaction affecting disease.

Analysis of the Interaction between Polygenic Risk Score and Calorie Intake in Obesity in the Korean Population.

INTRODUCTION: Obesity results from an imbalance in the intake and expenditure of calories that leads to lifestyle-related diseases. Although genome-wide association studies (GWAS) have revealed many obesity-related genetic factors, the interactions of these factors and calorie intake remain unknown. This study aimed to investigate interactions between calorie intake and the polygenic risk score (PRS) of BMI. METHODS: Three cohorts, i.e., from the Korea Association REsource (KARE; n = 8,736), CArdioVAscular Disease Association Study (CAVAS; n = 9,334), and Health EXAminee (HEXA; n = 28,445), were used for this study. BMI-related genetic loci were selected from previous GWAS. Two scores, PRS, and association (a)PRS, were used; the former was determined from 193 single-nucleotide polymorphisms (SNPs) from 5 GWAS datasets, and the latter from 62 SNPs (potentially associated) from 3 Korean cohorts (meta-analysis, p < 0.01). RESULTS: PRS and aPRS were significantly associated with BMI in all 3 cohorts but did not exhibit a significant interaction with total calorie intake. Similar results were obtained for obesity. PRS and aPRS were significantly associated with obesity but did not show a significant interaction with total calorie intake. We further analyzed the interaction with protein, fat, and carbohydrate intake. The results were similar to those for total calorie intake, with PRS and aPRS found to not be associated with the interaction of any of the 3 nutrition components for either BMI or obesity. DISCUSSION: The interaction of BMI PRS with calorie intake was investigated in 3 independent Korean cohorts (total n = 35,094) and no interactions were found between PRS and calorie intake for obesity.