Clinic Personalized Medicine Research Project Research Articles

Gwid: an R package and Shiny application for Genome-Wide analysis of IBD data.

Genome-wide identity by descent (gwid) is an R package developed for the analysis of identity-by-descent (IBD) data pertaining to dichotomous traits. This package offers a set of tools to assess differential IBD levels for the two states of a binary trait, yielding informative and meaningful results. Furthermore, it provides convenient functions to visualize the outcomes of these analyses, enhancing the interpretability and accessibility of the results. To assess the performance of the package, we conducted an evaluation using real genotype data derived from the SNPs to investigate rheumatoid arthritis susceptibility from the Marshfield Clinic Personalized Medicine Research Project. gwid is available as an open-source R package. Release versions can be accessed on CRAN (https://cran.r-project.org/package=gwid) for all major operating systems. The development version is maintained on GitHub (https://github.com/soroushmdg/gwid) and full documentation with examples and workflow templates is provided via the package website (http://tinyurl.com/gwid-tutorial). An interactive R Shiny dashboard is also developed (https://tinyurl.com/gwid-shiny).

A phenome-wide association study to discover pleiotropic effects of PCSK9, APOB, and LDLR

We conducted an electronic health record (EHR)-based phenome-wide association study (PheWAS) to discover pleiotropic effects of variants in three lipoprotein metabolism genes PCSK9, APOB, and LDLR. Using high-density genotype data, we tested the associations of variants in the three genes with 1232 EHR-derived binary phecodes in 51,700 European-ancestry (EA) individuals and 585 phecodes in 10,276 African-ancestry (AA) individuals; 457 PCSK9, 730 APOB, and 720 LDLR variants were filtered by imputation quality (r2 > 0.4), minor allele frequency (>1%), linkage disequilibrium (r2 < 0.3), and association with LDL-C levels, yielding a set of two PCSK9, three APOB, and five LDLR variants in EA but no variants in AA. Cases and controls were defined for each phecode using the PheWAS package in R. Logistic regression assuming an additive genetic model was used with adjustment for age, sex, and the first two principal components. Significant associations were tested in additional cohorts from Vanderbilt University (n = 29,713), the Marshfield Clinic Personalized Medicine Research Project (n = 9562), and UK Biobank (n = 408,455). We identified one PCSK9, two APOB, and two LDLR variants significantly associated with an examined phecode. Only one of the variants was associated with a non-lipid disease phecode, (“myopia”) but this association was not significant in the replication cohorts. In this large-scale PheWAS we did not find LDL-C-related variants in PCSK9, APOB, and LDLR to be associated with non-lipid-related phenotypes including diabetes, neurocognitive disorders, or cataracts.

Pollen Count and Presentation of Angiotensin-Converting Enzyme Inhibitor–Associated Angioedema

The incidence of angiotensin-converting enzyme (ACE) inhibitor-associated angioedema is increased in patients with seasonal allergies. We tested the hypothesis that patients with ACE inhibitor-associated angioedema present during months when pollen counts are increased. Cohort analysis examined the month of presentation of ACE inhibitor-associated angioedema and pollen counts in the ambulatory and hospital setting. Patients with ACE inhibitor-associated angioedema were ascertained through (1) an observational study of patients presenting to Vanderbilt University Medical Center, (2) patients presenting to the Marshfield Clinic and participating in the Marshfield Clinic Personalized Medicine Research Project, and (3) patients enrolled in The Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET). Measurements include date of presentation of ACE inhibitor-associated angioedema, population exposure to ACE inhibitor by date, and local pollen counts by date. At Vanderbilt, the rate of angioedema was significantly associated with tree pollen months (P = .01 from χ(2) test). When separate analyses were conducted in patients with a history of seasonal allergies and patients without, the rate of ACE inhibitor-associated angioedema was increased during tree pollen months only in patients with a history of seasonal allergies (P = .002). In Marshfield, the rate of angioedema was significantly associated with ragweed pollen months (P = .025). In ONTARGET, a positive trend was observed between the ACE inhibitor-associated angioedema rate and grass season, although it was not statistically significant (P = .057). Patients with ACE inhibitor-associated angioedema are more likely to present with this adverse drug event during months when pollen counts are increased.

Coronary risk assessment among intermediate risk patients using a clinical and biomarker based algorithm developed and validated in two population cohorts

Background:Many coronary heart disease (CHD) events occur in individuals classified as intermediate risk by commonly used assessment tools. Over half the individuals presenting with a severe cardiac event, such as myocardial infarction (MI), have at most one risk factor as included in the widely used Framingham risk assessment. Individuals classified as intermediate risk, who are actually at high risk, may not receive guideline recommended treatments. A clinically useful method for accurately predicting 5-year CHD risk among intermediate risk patients remains an unmet medical need.Objective:This study sought to develop a CHD Risk Assessment (CHDRA) model that improves 5-year risk stratification among intermediate risk individuals.Methods:Assay panels for biomarkers associated with atherosclerosis biology (inflammation, angiogenesis, apoptosis, chemotaxis, etc.) were optimized for measuring baseline serum samples from 1084 initially CHD-free Marshfield Clinic Personalized Medicine Research Project (PMRP) individuals. A multivariable Cox regression model was fit using the most powerful risk predictors within the clinical and protein variables identified by repeated cross-validation. The resulting CHDRA algorithm was validated in a Multiple-Ethnic Study of Atherosclerosis (MESA) case-cohort sample.Results:A CHDRA algorithm of age, sex, diabetes, and family history of MI, combined with serum levels of seven biomarkers (CTACK, Eotaxin, Fas Ligand, HGF, IL-16, MCP-3, and sFas) yielded a clinical net reclassification index of 42.7% (p < 0.001) for MESA patients with a recalibrated Framingham 5-year intermediate risk level. Across all patients, the model predicted acute coronary events (hazard ratio = 2.17, p < 0.001), and remained an independent predictor after Framingham risk factor adjustments.Limitations:These include the slightly different event definition with the MESA samples and inability to include PMRP fatal CHD events.Conclusions:A novel risk score of serum protein levels plus clinical risk factors, developed and validated in independent cohorts, demonstrated clinical utility for assessing the true risk of CHD events in intermediate risk patients. Improved accuracy in cardiovascular risk classification could lead to improved preventive care and fewer deaths.

Quantification of the Clinical Modifiers Impacting High-Density Lipoprotein Cholesterol in the Community: Personalized Medicine Research Project

High-density lipoprotein (HDL) cholesterol levels are inversely correlated with the development of cardiovascular disease. To date, genetic association studies have explained only a small proportion of the overall variance in HDL cholesterol. Further studies are needed, within practice-based cohorts, to place genetic findings into context alongside important clinical variables (eg, age, sex, body mass index, medication use, and clinical comorbidity). The Marshfield Clinic Personalized Medicine Research Project database was designed for large-scale studies of genetic epidemiology in a clinical practice-based setting. Because of its size and its unique practice-based design, this resource will provide adequate statistical power for the assessment of genetic findings related to HDL cholesterol level within the context of covariates known to modify lipid homeostasis. The authors report construction and validation of novel electronic phenotyping algorithms that can be used to model individual baseline HDL cholesterol levels within this practice-based resource. Because these algorithms were developed in a setting that reflects routine clinical care, future genetic studies using these algorithms within practice-based DNA biobanks should facilitate the identification of markers with optimal effect size after adjustment for known clinical factors contributing to the overall variance in HDL cholesterol level within the community.

Genetic Variation in CYP27B1 is Associated with Congestive Heart Failure in Patients with Hypertension

We tested the hypothesis that genetic variation in vitamin D-dependent signaling is associated with congestive heart failure in human subjects with hypertension. Functional polymorphisms were selected from five candidate genes: CYP27B1, CYP24A1, VDR, REN and ACE. Using the Marshfield Clinic Personalized Medicine Research Project, we genotyped 205 subjects with hypertension and congestive heart failure, 206 subjects with hypertension alone and 206 controls (frequency matched by age and gender). In the context of hypertension, a SNP in CYP27B1 was associated with congestive heart failure (odds ratio: 2.14 for subjects homozygous for the C allele; 95% CI: 1.05-4.39). Genetic variation in vitamin D biosynthesis is associated with increased risk of heart failure.

Estrogen receptor genotype is associated with risk of venous thromboembolism during tamoxifen therapy

Thromboembolism is a serious complication of tamoxifen therapy in women with breast cancer. Banked DNA from tamoxifen-treated individuals with breast cancer from the Marshfield Clinic Personalized Medicine Research Project, a population-based DNA repository, was tested for association between incidence of tamoxifen-associated thromboembolic events (TTE) and single nucleotide polymorphisms encoding the estrogen receptors 1,2 (ESR1, ESR2) or drug metabolism enzymes cytochrome P450 2D6 (CYP2D6) and aromatase (CYP19). TTE were experienced by 16/220 subjects with risk association noted for XbaI (rs9340799) genotype and ESR1 Xbal/PvuII diplotype (rs9340799 and rs2234693) (hazard ratio 3.47, 95% CI 0.97-12.44, P = 0.035). Association persisted after adjusting for classical risk factors including age at diagnosis and body mass index at enrollment. Initial evidence of association between increased risk for TTE and ESR1 genotype and ESR1 diplotype is presented. Determination of estrogen receptor genotype may identify a subset of women at increased risk for thromboembolism with tamoxifen exposure.

Community consultation and communication for a population‐based DNA biobank: The Marshfield clinic personalized medicine research project

The purpose of this article is to describe community consultation and communication efforts for the Personalized Medicine Research Project (PMRP), a population-based biobank. A series of focus group discussions was held in the year preceding initial recruitment efforts with potentially eligible community residents and slightly less than a year after initial recruitment with eligible residents who had declined participation in PMRP. A Community Advisory Group, with 19 members reflecting the demographics of the eligible community, was formed and meets twice yearly to provide advice and feedback to the PMRP Principal Investigator and the local IRB. Ongoing communication with study subjects, who consent on the condition that personal genetic results will not be disclosed, takes place through a newsletter that is distributed twice yearly, community talks and media coverage. Most focus group participants were concerned about the confidentiality of both their medical and genetic data. Focus group discussions with eligible residents who elected not to participate in PMRP revealed that many knew very little about the project, but thought that too much information had been provided, leading them to believe that it would take too long for them to understand and enroll in the study. In conclusion, an engaged community advisory group can provide a sounding board to study investigators for many study issues and can provide guidance for broader communication activities. Researchers need to balance the provision of information for potential subjects to make informed decisions about study participation, with respect for individuals' time to read and interpret study materials.

Characterization of Low‐Density Lipoprotein Cholesterol‐Lowering Efficacy for Atorvastatin in a Population‐Based DNA Biorepository

The Marshfield Clinic Personalized Medicine Research Project (PMRP) represents a large population-based biobank located in central Wisconsin. To position the PMRP database for large-scale pharmacogenetic association studies in the context of lipid-lowering therapy, we constructed an electronic phenotyping algorithm to quantify exposure and dose-response for atorvastatin, the most commonly prescribed lipid-lowering agent within this population. The resulting datasets were used to generate five distinct parameters for atorvastatin-induced changes in low-density lipoprotein (LDL) cholesterol level: (i) pretreatment, baseline LDL level [E(0)]; (ii) absolute reduction in LDL; (iii) relative reduction in LDL; (iv) potency [ED(50)]; and (v) maximal efficacy [E(max)]. These parameters will facilitate the efficient application of electronic phenotyping for drug outcomes in the context of large pharmacogenetic association studies.

The Marshfield Clinic Personalized Medicine Research Project: 2008 scientific update and lessons learned in the first 6years.

The Marshfield Clinic Personalized Medicine Research Project is the largest population-based biobank in the USA, with the ability to recontact subjects to obtain additional information to facilitate gene-environment studies. Nearly 20,000 adults have enrolled in the Personalized Medicine Research Project since 2001, after providing active written consent to access their Marshfield Clinic medical records to define phenotype and providing blood samples from which DNA, plasma and serum samples were stored. Numerous studies are underway in the area of pharmacogenetics and genetic epidemiology. In addition to the scientific discoveries being made, much has been learned regarding biobanking and the management of large amounts of data being generated. The purpose of this paper is to share the advice provided by the external Scientific Advisory Board and the scientific lessons learned along the way to build this research infrastructure and facilitate its use.

Genetic epidemiology studies of cataract in a population‐based biobank

Abstract Purpose To describe genetic epidemiology studies of cataract in the Marshfield Clinic Personalized Medicine Research Project (PMRP) Methods The PMRP is the largest population‐based biobank in the US. Nearly 20,000 adults aged 18 years and older have participated, with stored DNA, plasma and serum samples, and access to the Marshfield Clinic medical records. Subjects complete a brief questionniare that includes questions about smoking and alcohol intake. Height and weight are measured and body mass index is calculated. A candidate gene study of age‐related cataract has been completed and a genome‐wide association study (GWAS) of age‐related cataract is currently underway. Cataract cases are identified through diagnostic codes for cataract and procedure codes for lens extraction. Exclusionary diagnoses and procedures include congenital cataract, after‐cataract, traumatic cataract and steroid‐induced cataract. Controls are identified by the absence of any of the inclusion diagnoses and procedures in the context of an eye exam in subjects aged 50 years and older in the previous five years. Cataract location (cortical, nuclear, PSC) is identified through manual chart abstraction and natural language processing of dictated notes. Results In the candidate gene study of 1618 cataract surgery cases and 495 controls aged 60+, a SNP in the CRYGB gene proposed for study by Dr. Jochen Gray was found to be significantly associated with cataract surgery for females only. Electronic algorithms to accurately identify cases and controls for the GWAS study will be presented. Conclusion It is possible to conduct scientifically rigorous genetic epdiemiology studies of cataract in the context of a group medical practice. Care needs to be taken to accurately classify case and control status.

Intraocular Pressure Response to Medication in a Clinical Setting: The Marshfield Clinic Personalized Medicine Research Project

To estimate glaucoma and ocular hypertension prevalence and to describe temporal trends in prescribing patterns and intraocular pressure (IOP) response to topical medications used in glaucoma and ocular hypertension. The medical records of adult subjects enrolled in the population-based Marshfield Clinic Personalized Medicine Research Project were searched to identify participants who had been diagnosed with ocular hypertension or glaucoma and prescribed agent(s) to lower IOP. All IOPs before and after prescription of the IOP agents were recorded. As of December 31, 2005, 18,773 adults were enrolled in the Personalized Medicine Research Project, 57.1% were female, and their mean age was 50.3 years (range, 18 to 101 y). The overall rate of definite glaucoma in subjects aged 50 years and above was 2.1% (95% confidence interval=1.2, 2.4) and the rate of treated ocular hypertension was 1.4% (95% confidence interval=1.2, 1.7). Topical beta-blockers were the agents prescribed for the majority of subjects until the year 2000, when prostaglandins, first used in 1995, became the primary agent prescribed. In 2005, 75% of subjects used prostaglandin analogs and 46% used topical beta-blockers. The largest relative reduction in IOP in the first 3 months after prescription was observed for prostaglandin analogs (21.4% mean relative reduction), followed by beta-blockers (20.9% mean relative reduction). There has been a significant decrease over time in mean IOP before initiating medical therapy (linear regression beta coefficient=-0.30, P<0.0001, r=0.09). In this clinic-based setting, we found that treatment of glaucoma has changed over the past 20 years, with ophthalmologists more likely to begin treatment at lower baseline levels of IOP, and prostaglandin analogs the most commonly prescribed and agent to lower IOP.

Atorvastatin Dose‐response for High‐density Lipoprotein Cholesterol in a Population‐Based DNA Biobank Cohort

Statins have variable effects on high density lipoprotein (HDL) cholesterol. Prior data suggest that the atorvastatin dose‐response may be biphasic. The current study tests this hypothesis retrospectively, in a population‐based cohort. The Marshfield Clinic Personalized Medicine Research Project includes &gt;19,000 participants. Using natural language processing (NLP) software, we linked drug dosage to fasting HDL cholesterol levels for 2432 PMRP participants exposed to atorvastatin during the course of their routine clinical care. Mean HDL level prior to atorvastatin exposure was 46.2 mg/dl (S.D. = 11.9, N=2432). Mean baseline HDL of females (Mean=50.2, S.D. = 12.2, N=1256) was higher than males (Mean =41.9, S.D. = 10.0, N=1176). The lowest initial dose (5 mg daily) resulted in an increase in mean HDL to 51.01 mg/dl (male=46.3, female=55.4). Beyond 5 mg, however, each subsequent doubling of the daily dose resulted in a linear decrease in the mean HDL level. This biphasic dose‐response was noted in both genders. At the maximum daily dose of 80 mg, mean HDL was 46.62 mg/dl (male=43.8, female=50.2). Funding source U01HL069757‐06

Healthy People 2010 disease prevalence in the Marshfield Clinic Personalized Medicine Research Project cohort: opportunities for public health genomic research.

The purpose of this study was to estimate the prevalence of Healthy People2010 disease conditions in a large population-based cohort in central Wisconsin (WI, USA) and to consider how these conditions can be prioritized for research based on the use of healthcare services, the prevalence of various disease states and the resulting study power. Healthy People2010 diagnoses were estimated for participants in the Personalized Medicine Research Project (PMRP), a large population-based biobank for residents aged 18years and older living in central Wisconsin. By interrogating the electronic medical record, three parameters were calculated for each diagnosis: mean number of concomitant diagnoses, mean number of annual clinic visits before diagnosis and mean number of clinic visits after diagnosis. A total of 18,239adults enrolled in PMRP from September 2002 to May 2005 and were included in the study. They had a mean age of 49years (standard deviation:18.5), ranging from 18-98years; 57% were female. At least one Healthy People2010 disease was diagnosed in 86.4% of the participants; 13.6% had never been diagnosed with any of these conditions. The median number of diagnoses per subject was three (range:1-15). The median number of annual visits after diagnosis was lowest for chronic obstructive pulmonary disease (9.1) and highest for sleep apnea (17.9). Subjects with a diabetic retinopathy diagnosis had the highest number of concomitant diagnoses (mean:6.8). All of the diseases within the Healthy People2010 list are purported to have at least some genetic component, with the exception of injuries. The PMRP cohort is large enough that diseases of public health importance can be studied in the context of a variety of clinical and environmental covariates. This database is being developed as a national resource and is particularly useful where the estimated disease prevalence is 5% or greater. For less common diseases, additional cases can be recruited from throughout the Marshfield Clinic system of care, with population-based controls selected from the main PMRP study cohort.

Construction of Atorvastatin Dose–Response Relationships Using Data from a Large Population‐Based DNA Biobank

Construction of Atorvastatin Dose–Response Relationships Using Data from a Large Population‐Based DNA Biobank

Informed Consent and Subject Motivation to Participate in a Large, Population-Based Genomics Study: The Marshfield Clinic Personalized Medicine Research Project

Background: The objective of this study was to measure subject perspective and reaction to participation in the Personalized Medicine Research Project (PMRP) and to identify factors predicting understanding of the study elements. Method: Self-administered questionnaires were mailed to 1,593 subjects (10% sample). The questionnaire had three sections: section A consisted of 21 factual questions; section B consisted of 14 questions to assess the level of understanding about the PMRP concepts, and section C asked about the purpose of the PMRP. Results: The mean age of the 924 survey respondents was 52 years (SD = 16.9), with a range of 18–95 years. The majority of participants were female (n = 561, 61%). The percent of total correct responses for section A was significantly higher for females compared with males (males: 58.4% and females: 60.4%, t test = –2.18, p = 0.03) and age was significantly inversely related to percent of correct responses (β coefficient = –0.122, p < 0.001). More than one third of the participants indicated that the USD 20 greatly influenced their decision to participate in the project. In a multiple logistic regression model, people living outside of Marshfield were significantly more likely to indicate that the USD 20 greatly influenced their decision to participate (odds ratio = 1.40, 95% confidence limit = 1.06, 1.86) and age was inversely related to the monetary influence on decision to participate (odds ratio = 0.98, 95% confidence limit = 0.97, 0.98). Conclusion: Future community consultation efforts should highlight areas of lower understanding. In addition, research coordinators may need to take more time informing males and older individuals about project details so that they are making truly informed decisions about study participation.

Statin‐use phenotyping in the advanced elderly ‐ validation in a large, population‐based DNA Biobank

Hydroxy-methyl-glutaryl (HMG) coenzyme A reductase inhibitors (statins) are considered both safe and efficacious. To quantify statin use in the advanced elderly, a cohort of subjects over 80 years of age was identified within the Marshfield Clinic Personalized Medicine Research Project (PMRP). The PMRP represents one of the largest population-based DNA Biobanks in the world (Science 298 1158–1161, 2002). Ninety-five subjects were identified (mean age 88.3 + 3.2 years), and retrospective drug-use histories were constructed through manual data abstraction. Twenty-one percent of subjects were using a statin at most recent examination (95% C.I. 13.4 – 30.6). To assess the utility of electronic phenotyping, we constructed independent statin use histories in the same subjects through the application of natural language processing software (Language and Computing, Inc.). The electronic phenotype for current medication use was 100% sensitive and 96% specific. The initial positive predictive value (PPV) was 87%. After programmatic selection of only those electronic free text data containing a dose, PPV improved to 95%. Electronic statin use histories were then constructed for a 10-year period, and collated with electronically abstracted lipid panels, for each respective subject. This phenotyping strategy should facilitate characterization of electronic dose response curves in large population-based cohorts, and position investigators to conduct pharmacogenetic association studies on an unprecedented scale. (This work was funded by a grant from the Marshfield Clinic Research Foundation.)

432: Informed Consent and Subject Motivation to Participate in a Large, Population-Based Genomics Study: The Marshfield Clinic Personalized Medicine Research Project

432: Informed Consent and Subject Motivation to Participate in a Large, Population-Based Genomics Study: The Marshfield Clinic Personalized Medicine Research Project

Marshfield Clinic Personalized Medicine Research Project (PMRP): design, methods and recruitment for a large population-based biobank.

The objective of this paper is to summarize the planning for Phase I of the Marshfield Clinic Personalized Medicine Research Project (PMRP) and to describe the recruitment efforts in the first 2 years. The purpose of Phase I of the PMRP was to develop a large population-based biobank with DNA, plasma and serum samples to facilitate genomics research. Planning and consultation was facilitated with three external boards: the Ethics and Security Advisory Board; the Scientific Advisory Board; and the Community Advisory Group. Commencing in September 2002, residents aged 18 and above who resided in 1 of 19 zip codes surrounding Marshfield, WI, USA, were invited to participate. After providing written informed consent, participants completed brief questionnaires that included questions about demographics, some environmental exposures, family history of disease, and adverse drug reactions, as well as family members living in the study area. Participants provided 50ml of blood from which DNA was extracted and plasma and serum samples were stored. The informed consent document allowed access to electronic medical records and included language about non-disclosure of personal research results. A tick-off box was also included so that participants could either allow or decline subsequent recontact for future research studies. A total of 17,463 subjects were enrolled during the first 23months of recruitment (44.3% of the residents who the Research Project Assistants were able to contact). The participants ranged in age from 18 to 98.5years (mean = 48.9, median = 48); 57.2% (n = 9986) were female. Self-reported race in the study cohort was similar to the year 2000 census for Wood County, WI, USA, with the majority (98%) reporting themselves to be White Caucasian. The majority of subjects (n= 13,391, 76.7%) indicated that they had German ancestry. Only 142participants (< 1%) opted out on the consent form for contact for future studies. The majority of the cohort reported that their current area of residence was a suburb, city or village (n = 10630, 60.87%); the remainder reported residence in a rural home or hobby farm (n=5365, 30.72%), or a working farm or ranch (n = 1451, 8.31%). More than half the cohort (n = 9409, 53.88%) had lived on a working farm at some point in their life. The PMRP database will allow research in three areas: genetic epidemiology, pharmacogenetics, and population genetics. The size and the stability of the population as well as the relative ethnic homogeneity will help facilitate longitudinal studies with valid research results that are not biased by population stratification.