Identification of Pancreatic Cancer Germline Risk Variants With Effects That Are Modified by Smoking.

PURPOSE: Pancreatic cancer (PC) is a deadly disease most often diagnosed in late stages. Identification of high-risk subjects could both contribute to preventative measures and help diagnose the disease at earlier timepoints. However, known risk factors, assessed independently, are currently insufficient for accurately stratifying patients. We use large-scale data from the UK Biobank (UKB) to identify genetic variant-smoking interaction effects and show their importance in risk assessment. METHODS: We draw data from 15,086,830 genetic variants and 315,512 individuals in the UKB. There are 765 cases of PC. Crucially, robust resampling corrections are used to overcome well-known challenges in hypothesis testing for interactions. Replication analysis is conducted in two independent cohorts totaling 793 cases and 570 controls. Integration of functional annotation data and construction of polygenic risk scores (PRS) demonstrate the additional insight provided by interaction effects. RESULTS: We identify the genome-wide significant variant rs77196339 on chromosome 2 (per minor allele odds ratio in never-smokers, 2.31 [95% CI, 1.69 to 3.15]; per minor allele odds ratio in ever-smokers, 0.53 [95% CI, 0.30 to 0.91]; P = 3.54 × 10-8) as well as eight other loci with suggestive evidence of interaction effects (P < 5 × 10-6). The rs77196339 region association is validated (P < .05) in the replication sample. PRS incorporating interaction effects show improved discriminatory ability over PRS of main effects alone. CONCLUSION: This study of genome-wide germline variants identified smoking to modify the effect of rs77196339 on PC risk. Interactions between known risk factors can provide critical information for identifying high-risk subjects, given the relative inadequacy of models considering only main effects, as demonstrated in PRS. Further studies are necessary to advance toward comprehensive risk prediction approaches for PC.

Trans-Ancestral Genetic Risk Factors for Treatment-Related Type 2 Diabetes Mellitus in Survivors of Childhood Cancer.

PURPOSE: Type 2 diabetes mellitus (T2D) is a prevalent long-term complication of treatment in survivors of childhood cancer, with marked racial/ethnic differences in burden. In this study, we investigated trans-ancestral genetic risks for treatment-related T2D. PATIENTS AND METHODS: Leveraging whole-genome sequencing data from the St Jude Lifetime Cohort (N = 3,676, 304 clinically ascertained cases), we conducted ancestry-specific genome-wide association studies among survivors of African and European genetic ancestry (AFR and EUR, respectively) followed by trans-ancestry meta-analysis. Trans-/within-ancestry replication including data from the Childhood Cancer Survivor Study (N = 5,965) was required for prioritization. Three external general population T2D polygenic risk scores (PRSs) were assessed, including multiancestry PRSs. Treatment risk effect modification was evaluated for prioritized loci. RESULTS: Four novel T2D risk loci showing trans-/within-ancestry replication evidence were identified, with three loci achieving genome-wide significance (P < 5 × 10-8). Among these, common variants at 5p15.2 (LINC02112), 2p25.3 (MYT1L), and 19p12 (ZNF492) showed evidence of modifying alkylating agent-related T2D risk in both ancestral groups, but showed disproportionately greater risk in AFR survivors (AFR odds ratios [ORs], 3.95-17.81; EUR ORs, 2.37-3.32). In survivor-specific RNA-sequencing data (N = 207), the 19p12 locus variant was associated with greater ZNF492 expression dysregulation after exposures to alkylators. Elevated T2D risks across ancestry groups were only observed with increasing values for multiancestry T2D PRSs and were especially increased among survivors treated with alkylators (top v bottom quintiles: ORAFR, 20.18; P = .023; OREUR, 13.44; P = 1.3 × 10-9). CONCLUSION: Our findings suggest therapy-related genetic risks contribute to the increased T2D burden among non-Hispanic Black childhood cancer survivors. Additional study of how therapy-related genetic susceptibility contributes to this disparity is needed.

Inferring a directed acyclic graph of phenotypes from GWAS summary statistics.

Estimating phenotype networks is a growing field in computational biology. It deepens the understanding of disease etiology and is useful in many applications. In this study, we present a method that constructs a phenotype network by assuming a Gaussian linear structure model embedding a directed acyclic graph (DAG). We utilize genetic variants as instrumental variables and show how our method only requires access to summary statistics from a genome-wide association study (GWAS) and a reference panel of genotype data. Besides estimation, a distinct feature of the method is its summary statistics-based likelihood ratio test on directed edges. We applied our method to estimate a causal network of 29 cardiovascular-related proteins and linked the estimated network to Alzheimer's disease (AD). A simulation study was conducted to demonstrate the effectiveness of this method. An R package sumdag implementing the proposed method, all relevant code, and a Shiny application are available.

Subset scanning for multi-trait analysis using GWAS summary statistics.

MOTIVATION: Multi-trait analysis has been shown to have greater statistical power than single-trait analysis. Most of the existing multi-trait analysis methods only work with a limited number of traits and usually prioritize high statistical power over identifying relevant traits, which heavily rely on domain knowledge. RESULTS: To handle diseases and traits with obscure etiology, we developed TraitScan, a powerful and fast algorithm that identifies potential pleiotropic traits from a moderate or large number of traits (e.g. dozens to thousands) and tests the association between one genetic variant and the selected traits. TraitScan can handle either individual-level or summary-level GWAS data. We evaluated TraitScan using extensive simulations and found that it outperformed existing methods in terms of both testing power and trait selection when sparsity was low or modest. We then applied it to search for traits associated with Ewing Sarcoma, a rare bone tumor with peak onset in adolescence, among 754 traits in UK Biobank. Our analysis revealed a few promising traits worthy of further investigation, highlighting the use of TraitScan for more effective multi-trait analysis as biobanks emerge. We also extended TraitScan to search and test association with a polygenic risk score and genetically imputed gene expression. AVAILABILITY AND IMPLEMENTATION: Our algorithm is implemented in an R package "TraitScan" available at https://github.com/RuiCao34/TraitScan.

Genetic analyses identify evidence for a causal relationship between Ewing sarcoma and hernias.

Knowledge of Ewing sarcoma (EWS) risk factors is exceedingly limited; however, multiple small, independent studies have suggested a possible connection between hernia and EWS. By leveraging hernia summary statistics from the UK Biobank and a recently published genome-wide association study of EWS (733 EWS cases and 1,346 controls), we conducted a genetic investigation of the relationship of 5 hernia types (diaphragmatic, inguinal, umbilical, femoral, and ventral) and EWS. We discovered a positive causal relationship between inguinal hernia and EWS (OR 1.27, 95% confidence interval [CI] 1.01-1.59, and p = 0.041) through Mendelian randomization analysis. Further analyses suggested shared pathways through three genes: HMGA2, LOX, and FBXW7. Diaphragmatic hernia showed a stronger causal relationship with EWS among all of the hernia types (OR 2.26, 95% CI 1.30-3.95, p = 0.004), but no statistically significant local correlation pattern was observed. No evidence of a causal or genetic relationship was observed between EWS and the other three hernia types, including umbilical hernia, despite a previous report indicating an OR as high as 3.3. The finding of our genetic analysis provided additional support to the hypothesis that EWS and hernias may share a common origin.

Inferring a directed acyclic graph of phenotypes from GWAS summary statistics.

Estimating phenotype networks is a growing field in computational biology. It deepens the understanding of disease etiology and is useful in many applications. In this study, we present a method that constructs a phenotype network by assuming a Gaussian linear structure model embedding a directed acyclic graph (DAG). We utilize genetic variants as instrumental variables and show how our method only requires access to summary statistics from a genome-wide association study (GWAS) and a reference panel of genotype data. Besides estimation, a distinct feature of the method is its summary statistics-based likelihood ratio test on directed edges. We applied our method to estimate a causal network of 29 cardiovascular-related proteins and linked the estimated network to Alzheimer's disease (AD). A simulation study was conducted to demonstrate the effectiveness of this method. An R package sumdag implementing the proposed method, all relevant code, and a Shiny application are available at https://github.com/chunlinli/sumdag.

Genetically inferred birthweight, height, and puberty timing and risk of osteosarcoma.

INTRODUCTION: Several studies have linked increased risk of osteosarcoma with tall stature, high birthweight, and early puberty, although evidence is inconsistent. We used genetic risk scores (GRS) based on established genetic loci for these traits and evaluated associations between genetically inferred birthweight, height, and puberty timing with osteosarcoma. METHODS: Using genotype data from two genome-wide association studies, totaling 1039 cases and 2923 controls of European ancestry, association analyses were conducted using logistic regression for each study and meta-analyzed to estimate pooled odds ratios (ORs) and 95% confidence intervals (CIs). Subgroup analyses were conducted by case diagnosis age, metastasis status, tumor location, tumor histology, and presence of a known pathogenic variant in a cancer susceptibility gene. RESULTS: Genetically inferred higher birthweight was associated with an increased risk of osteosarcoma (OR =1.59, 95% CI 1.07-2.38, P = 0.02). This association was strongest in cases without metastatic disease (OR =2.46, 95% CI 1.44-4.19, P = 9.5 ×10-04). Although there was no overall association between osteosarcoma and genetically inferred taller stature (OR=1.06, 95% CI 0.96-1.17, P = 0.28), the GRS for taller stature was associated with an increased risk of osteosarcoma in 154 cases with a known pathogenic cancer susceptibility gene variant (OR=1.29, 95% CI 1.03-1.63, P = 0.03). There were no significant associations between the GRS for puberty timing and osteosarcoma. CONCLUSION: A genetic propensity to higher birthweight was associated with increased osteosarcoma risk, suggesting that shared genetic factors or biological pathways that affect birthweight may contribute to osteosarcoma pathogenesis.

Speeding up interval estimation for R2-based mediation effect of high-dimensional mediators via cross-fitting.

Mediation analysis is a useful tool in biomedical research to investigate how molecular phenotypes, such as gene expression, mediate the effect of an exposure on health outcomes. However, commonly used mean-based total mediation effect measures may suffer from cancellation of component-wise mediation effects of opposite directions in the presence of high-dimensional omics mediators. To overcome this limitation, a variance-based R-squared total mediation effect measure has been recently proposed, which, nevertheless, relies on the computationally intensive nonparametric bootstrap for confidence interval estimation. In this work, we formulate a more efficient two-stage cross-fitted estimation procedure for the R-squared measure. To avoid potential bias, we perform iterative Sure Independence Screening (iSIS) in two subsamples to exclude the non-mediators, followed by ordinary least squares (OLS) regressions for the variance estimation. We then construct confidence intervals based on the newly-derived closed-form asymptotic distribution of the R-squared measure. Extensive simulation studies demonstrate that the proposed procedure is hundreds of times more computationally efficient than the resampling-based method with comparable coverage probability. Furthermore, when applied to the Framingham Heart Study, the proposed method replicated the established finding of gene expression mediating age-related variation in systolic blood pressure and discovered the role of gene expression profiles in the relationship between sex and high-density lipoprotein cholesterol. The proposed cross-fitted interval estimation procedure is implemented in R package RsqMed.

Predicted leukocyte telomere length and risk of germ cell tumours.

BACKGROUND: Genetically predicted leukocyte telomere length (LTL) has been evaluated in several studies of childhood and adult cancer. We test whether genetically predicted longer LTL is associated with germ cell tumours (GCT) in children and adults. METHODS: Paediatric GCT samples were obtained from a Children's Oncology Group study and state biobank programs in California and Michigan (N = 1413 cases, 1220 biological parents and 1022 unrelated controls). Replication analysis included 396 adult testicular GCTs (TGCT) and 1589 matched controls from the UK Biobank. Mendelian randomisation was used to look at the association between genetically predicted LTL and GCTs and TERT variants were evaluated within GCT subgroups. RESULTS: We identified significant associations between TERT variants reported in previous adult TGCT GWAS in paediatric GCT: TERT/rs2736100-C (OR = 0.82; P = 0.0003), TERT/rs2853677-G (OR = 0.80; P = 0.001), and TERT/rs7705526-A (OR = 0.81; P = 0.003). We also extended these findings to females and tumours outside the testes. In contrast, we did not observe strong evidence for an association between genetically predicted LTL by other variants and GCT risk in children or adults. CONCLUSION: While TERT is a known susceptibility locus for GCT, our results suggest that LTL predicted by other variants is not strongly associated with risk in either children or adults.

Impact of fetal expression quantitative trait loci on transcriptome-wide association study of childhood leukemia.

Transcriptome-wide association studies increase the yield of loci associated with disease phenotypes by focusing on expression quantitative trait loci (eQTL). The major source of eQTL data for is the Gene and Tissue Expression (GTEx) project, which is comprised entirely of adults, mainly those >50 years of age at death. Since gene expression levels differ by developmental stage, it is not clear whether eQTLs derived from adult data sources are best suited for use in young-onset diseases such as pediatric cancers. To fill in this knowledge gap, we performed a large-scale eQTL mapping analysis in the GenCord study with newborn samples and compared it with GTEx. Under matched conditions, we found around 80% of the eQTLs in one study can be replicated in the other. However, among all eQTLs identified in GenCord (GTEx), 584 (1045) showed statistically significant differences in effect sizes in GTEx (GenCord). We further investigated how using fetal eQTL data can facilitate the genetic association study of acute lymphoblastic leukemia. GenCord and GTEx identified the same genetic loci with statistical significance; however, the overall association pattern was only weakly correlated. Our paper demonstrates age-differential eQTLs and shows their potential influence on childhood leukemia research.

Rare and low-frequency exonic variants and gene-by-smoking interactions in pulmonary function.

Genome-wide association studies have identified numerous common genetic variants associated with spirometric measures of pulmonary function, including forced expiratory volume in one second (FEV1), forced vital capacity, and their ratio. However, variants with lower minor allele frequencies are less explored. We conducted a large-scale gene-smoking interaction meta-analysis on exonic rare and low-frequency variants involving 44,429 individuals of European ancestry in the discovery stage and sought replication in the UK BiLEVE study with 45,133 European ancestry samples and UK Biobank study with 59,478 samples. We leveraged data on cigarette smoking, the major environmental risk factor for reduced lung function, by testing gene-by-smoking interaction effects only and simultaneously testing the genetic main effects and interaction effects. The most statistically significant signal that replicated was a previously reported low-frequency signal in GPR126, distinct from common variant associations in this gene. Although only nominal replication was obtained for a top rare variant signal rs142935352 in one of the two studies, interaction and joint tests for current smoking and PDE3B were significantly associated with FEV1. This study investigates the utility of assessing gene-by-smoking interactions and underscores their effects on potential pulmonary function.

Estimation of total mediation effect for high-dimensional omics mediators.

BACKGROUND: Environmental exposures can regulate intermediate molecular phenotypes, such as gene expression, by different mechanisms and thereby lead to various health outcomes. It is of significant scientific interest to unravel the role of potentially high-dimensional intermediate phenotypes in the relationship between environmental exposure and traits. Mediation analysis is an important tool for investigating such relationships. However, it has mainly focused on low-dimensional settings, and there is a lack of a good measure of the total mediation effect. Here, we extend an R-squared (R[Formula: see text]) effect size measure, originally proposed in the single-mediator setting, to the moderate- and high-dimensional mediator settings in the mixed model framework. RESULTS: Based on extensive simulations, we compare our measure and estimation procedure with several frequently used mediation measures, including product, proportion, and ratio measures. Our R[Formula: see text]-based second-moment measure has small bias and variance under the correctly specified model. To mitigate potential bias induced by non-mediators, we examine two variable selection procedures, i.e., iterative sure independence screening and false discovery rate control, to exclude the non-mediators. We establish the consistency of the proposed estimation procedures and introduce a resampling-based confidence interval. By applying the proposed estimation procedure, we found that 38% of the age-related variations in systolic blood pressure can be explained by gene expression profiles in the Framingham Heart Study of 1711 individuals. An R package "RsqMed" is available on CRAN. CONCLUSION: R-squared (R[Formula: see text]) is an effective and efficient measure for total mediation effect especially under high-dimensional setting.

Integrative analysis of multi-omics data for discovering low-frequency variants associated with low-density lipoprotein cholesterol levels.

MOTIVATION: The abundance of omics data has facilitated integrative analyses of single and multiple molecular layers with genome-wide association studies focusing on common variants. Built on its successes, we propose a general analysis framework to leverage multi-omics data with sequencing data to improve the statistical power of discovering new associations and understanding of the disease susceptibility due to low-frequency variants. The proposed test features its robustness to model misspecification, high power across a wide range of scenarios and the potential of offering insights into the underlying genetic architecture and disease mechanisms. RESULTS: Using the Framingham Heart Study data, we show that low-frequency variants are predictive of DNA methylation, even after conditioning on the nearby common variants. In addition, DNA methylation and gene expression provide complementary information to functional genomics. In the Avon Longitudinal Study of Parents and Children with a sample size of 1497, one gene CLPTM1 is identified to be associated with low-density lipoprotein cholesterol levels by the proposed powerful adaptive gene-based test integrating information from gene expression, methylation and enhancer-promoter interactions. It is further replicated in the TwinsUK study with 1706 samples. The signal is driven by both low-frequency and common variants. AVAILABILITY AND IMPLEMENTATION: Models are available at https://github.com/ytzhong/DNAm. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Incorporating multiple sets of eQTL weights into gene-by-environment interaction analysis identifies novel susceptibility loci for pancreatic cancer.

It is of great scientific interest to identify interactions between genetic variants and environmental exposures that may modify the risk of complex diseases. However, larger sample sizes are usually required to detect gene-by-environment interaction (G × E) than required to detect genetic main association effects. To boost the statistical power and improve the understanding of the underlying molecular mechanisms, we incorporate functional genomics information, specifically, expression quantitative trait loci (eQTLs), into a data-adaptive G × E test, called aGEw. This test adaptively chooses the best eQTL weights from multiple tissues and provides an extra layer of weighting at the genetic variant level. Extensive simulations show that the aGEw test can control the Type 1 error rate, and the power is resilient to the inclusion of neutral variants and noninformative external weights. We applied the proposed aGEw test to the Pancreatic Cancer Case-Control Consortium (discovery cohort of 3,585 cases and 3,482 controls) and the PanScan II genome-wide association study data (replication cohort of 2,021 cases and 2,105 controls) with smoking as the exposure of interest. Two novel putative smoking-related pancreatic cancer susceptibility genes, TRIP10 and KDM3A, were identified. The aGEw test is implemented in an R package aGE.

Integrating DNA sequencing and transcriptomic data for association analyses of low-frequency variants and lipid traits.

Transcriptome-wide association studies (TWAS) integrate genome-wide association studies (GWAS) and transcriptomic data to showcase their improved statistical power of identifying gene-trait associations while, importantly, offering further biological insights. TWAS have thus far focused on common variants as available from GWAS. Compared with common variants, the findings for or even applications to low-frequency variants are limited and their underlying role in regulating gene expression is less clear. To fill this gap, we extend TWAS to integrating whole genome sequencing data with transcriptomic data for low-frequency variants. Using the data from the Framingham Heart Study, we demonstrate that low-frequency variants play an important and universal role in predicting gene expression, which is not completely due to linkage disequilibrium with the nearby common variants. By including low-frequency variants, in addition to common variants, we increase the predictivity of gene expression for 79% of the examined genes. Incorporating this piece of functional genomic information, we perform association testing for five lipid traits in two UK10K whole genome sequencing cohorts, hypothesizing that cis-expression quantitative trait loci, including low-frequency variants, are more likely to be trait-associated. We discover that two genes, LDLR and TTC22, are genome-wide significantly associated with low-density lipoprotein cholesterol based on 3203 subjects and that the association signals are largely independent of common variants. We further demonstrate that a joint analysis of both common and low-frequency variants identifies association signals that would be missed by testing on either common variants or low-frequency variants alone.

An adaptive test for meta-analysis of rare variant association studies.

Single genome-wide studies may be underpowered to detect trait-associated rare variants with moderate or weak effect sizes. As a viable alternative, meta-analysis is widely used to increase power by combining different studies. The power of meta-analysis critically depends on the underlying association patterns and heterogeneity levels, which are unknown and vary from locus to locus. However, existing methods mainly focus on one or only a few combinations of the association pattern and heterogeneity level, thus may lose power in many situations. To address this issue, we propose a general and unified framework by combining a class of tests including and beyond some existing ones, leading to high power across a wide range of scenarios. We demonstrate that the proposed test is more powerful than some existing methods in simulation studies, then show their performance with the NHLBI Exome-Sequencing Project (ESP) data. One gene (B4GALNT2) was found by our proposed test, but not by others, to be statistically significantly associated with plasma triglyceride. The signal was driven by African-ancestry subjects but it was previously reported to be associated with coronary artery disease among European-ancestry subjects. We implemented our method in an R package aSPUmeta, publicly available at https://github.com/ytzhong/metaRV and will be on CRAN soon.

Dynamic prediction of hospital admission with medical claim data.

BACKGROUND: Congestive heart failure is one of the most common reasons those aged 65 and over are hospitalized in the United States, which has caused a considerable economic burden. The precise prediction of hospitalization caused by congestive heart failure in the near future could prevent possible hospitalization, optimize the medical resources, and better meet the healthcare needs of patients. METHODS: To fully utilize the monthly-updated claim feed data released by The Centers for Medicare and Medicaid Services (CMS), we present a dynamic random survival forest model adapted for periodically updated data to predict the risk of adverse events. We apply our model to dynamically predict the risk of hospital admission among patients with congestive heart failure identified using the Accountable Care Organization Operational System Claim and Claim Line Feed data from Feb 2014 to Sep 2015. We benchmark the proposed model with two commonly used models in medical application literature: the cox proportional model and logistic regression model with L-1 norm penalty. RESULTS: Results show that our model has high Area-Under-the-ROC-Curve across time points and C-statistics. In addition to the high performance, it provides measures of variable importance and individual-level instant risk. CONCLUSION: We present an efficient model adapted for periodically updated data such as the monthly updated claim feed data released by CMS to predict the risk of hospitalization. In addition to processing big-volume periodically updated stream-like data, our model can capture event onset information and time-to-event information, incorporate time-varying features, provide insights of variable importance and have good prediction power. To the best of our knowledge, it is the first work combining sliding window technique with the random survival forest model. The model achieves remarkable performance and could be easily deployed to monitor patients in real time.

A powerful and data-adaptive test for rare-variant-based gene-environment interaction analysis.

As whole-exome/genome sequencing data become increasingly available in genetic epidemiology research consortia, there is emerging interest in testing the interactions between rare genetic variants and environmental exposures that modify the risk of complex diseases. However, testing rare-variant-based gene-by-environment interactions (GxE) is more challenging than testing the genetic main effects due to the difficulty in correctly estimating the latter under the null hypothesis of no GxE effects and the presence of neutral variants. In response, we have developed a family of powerful and data-adaptive GxE tests, called "aGE" tests, in the framework of the adaptive powered score test, originally proposed for testing the genetic main effects. Using extensive simulations, we show that aGE tests can control the type I error rate in the presence of a large number of neutral variants or a nonlinear environmental main effect, and the power is more resilient to the inclusion of neutral variants than that of existing methods. We demonstrate the performance of the proposed aGE tests using Pancreatic Cancer Case-Control Consortium Exome Chip data. An R package "aGE" is available at http://github.com/ytzhong/projects/.

Optically Unraveling the Edge Chirality-Dependent Band Structure and Plasmon Damping in Graphene Edges.

The nontrivial topological origin and pseudospinorial character of electron wavefunctions make edge states possess unusual electronic properties. Twenty years ago, the tight-binding model calculation predicted that zigzag termination of 2D sheets of carbon atoms have peculiar edge states, which show potential application in spintronics and modern information technologies. Although scanning probe microscopy is employed to capture this phenomenon, the experimental demonstration of its optical response remains challenging. Here, the propagating graphene plasmon provides an edge-selective polaritonic probe to directly detect and control the electronic edge state at ambient condition. Compared with armchair, the edge-band structure in the bandgap gives rise to additional optical absorption and strongly absorbed rim at zigzag edge. Furthermore, the optical conductivity is reconstructed and the anisotropic plasmon damping in graphene systems is revealed. The reported approach paves the way for detecting edge-specific phenomena in other van der Waals materials and topological insulators.

Caffeine, creatine, GRIN2A and Parkinson's disease progression.

Caffeine is neuroprotective in animal models of Parkinson's disease (PD) and caffeine intake is inversely associated with the risk of PD. This association may be influenced by the genotype of GRIN2A, which encodes an NMDA-glutamate-receptor subunit. In two placebo-controlled studies, we detected no association of caffeine intake with the rate of clinical progression of PD, except among subjects taking creatine, for whom higher caffeine intake was associated with more rapid progression. We now have analyzed data from 420 subjects for whom DNA samples and caffeine intake data were available from a placebo-controlled study of creatine in PD. The GRIN2A genotype was not associated with the rate of clinical progression of PD in the placebo group. However, there was a 4-way interaction between GRIN2A genotype, caffeine, creatine and the time since baseline. Among subjects in the creatine group with high levels of caffeine intake, but not among those with low caffeine intake, the GRIN2A T allele was associated with more rapid progression (p=0.03). These data indicate that the deleterious interaction between caffeine and creatine with respect to rate of progression of PD is influenced by GRIN2A genotype. This example of a genetic factor interacting with environmental factors illustrates the complexity of gene-environment interactions in the progression of PD.