Fall 2018 Seminars


irvinRyan Irvin, PhD, MS | View Seminar
Associate Professor, School of Public Health
University of Alabama, Birmingham

'The emerging role of epigenetics in cardiometabolic disease'

The high prevalence of metabolic disease among US adults has resulted in significant increases in cardiovascular diseases. These disorders disproportionately affect minority populations in the US. Identifying epigenetic alterations associated with metabolic traits has provided additional information regarding etiology beyond current evidence from Genome Wide Association Studies. Phenotypic data from observational epidemiology cohorts paired with DNA methylation data in large studies of multiple ethnicities have identified a handful of biologically plausible genes associated with fasting glucose, blood pressure, lipid levels, and body mass index. Often the same genes are associated with multiple traits across multiple racial groups. Since most studies are cross-sectional many questions remain as to whether epigenetic alterations in these genes are the cause or consequence of metabolic trait variation. Bi-directional Mendelian Randomization studies and animal studies have helped determine directionality of findings. Future studies are needed to understand if epigenetic markers are useful for disease prediction.


cadilhacDominique Cadilhac, PhD, MPH | View Seminar
Professor Translational Public Health and Evaluation Division 
Monash University

“10 years of the Australian Stroke Clinical Registry - the what, the how and the why”

Professor Cadilhac will provide an overview of clinical quality disease registries in Australia with a focus on the Australian Stroke Clincial Registry that was established in 2009 and is used by over 70 hospitals. A summary of how the data have been used to change clincial practice and contribute to the broader evidence base on stroke outcomes including quality of life will be provided. Innovation to maximise the use of the data for hospitals, policy decisions making and practice including data linkage will be highlighted.


bJohann Gagnon Bartsch, PhD, MS | View Seminar
Assistant Professor
University of Michigan

"The Duality of Negative Controls and Replicates"

Negative controls can be used to adjust for unobserved confounders in an observational study. A negative control is a variable that is known a priori to be (1) unaffected by treatment, and (2) affected by the unobserved confounders. Any observed variation in a negative control may be attributed to the confounders, but not to treatment. Thus, negative controls can be used to partially identify the unobserved confounders. A similar situation arises when a single observational unit is observed multiple times, under varying conditions of the confounders. The multiple observations are referred to as replicates. Any observed variation between the replicates may be attributed purely to the confounders. Thus, like negative controls, replicates can be used to partially identify the confounding variables. Importantly, in a high-dimensional setting, the partial identification provided by negative controls and the partial identification provided by the replicates are in some sense dual to one another. More to the point, these two partial identifications are not redundant, but rather complimentary, and therefore negative controls and replicates can be used together to more fully identify and control for unobserved confounders. In this talk, I will demonstrate the use of negative controls and replicates to remove batch effects and other unwanted variation from genomic data, including microarray, nanostring, and single-cell RNA sequencing data.


arnettDonna Arnett, PhD, MSPH | View Seminar
Dean and Professor, College of Public Health
University of Kentucky, Lexington
Past President, American Heart Association

"HyperGEN—Genetics of Left Ventricular Hypertrophy: Omic approaches to understanding echocardiographic traits"

Left ventricular (LV) mass and related echocardiographic phenotypes are heritable, and they are important predictors of cardiovascular disease, particularly in hypertensive individuals. The HyperGEN: Genetics of Left Ventricular Hypertrophy (LVH) Study has been using cutting-edge omic approaches in a cohort of hypertensive sibships to discover the genetic, genomic, and epigenomic correlates of these traits since 1995. As such, HyperGEN: LVH serves as excellent case study to illustrate evolving genetic epidemiological methods and the power, limitations, and challenges of these methods. We will trace the history of HyperGEN: LVH through the eras of linkage and candidate gene analysis, genome-wide association studies, whole-exome sequencing studies, and epigenome-wide association studies into the current era of whole-genome sequencing. The use of animal and cellular models, gene expression analysis, and pathway analysis in conjunction with association studies will also be discussed. 


mellorSidra Goldman Mellor, PhD, MPH | View Seminar
Assistant Professor, Public Health
University of California, Merced

"Self-harm behavior and health across the lifecourse: Insights from the emergency department"

Sidra Goldman-Mellor, Ph.D., received her doctoral degree in epidemiology from the University of California, Berkeley, and has been an Assistant Professor of Public Health at the University of California, Merced since 2014. Her NIH-funded research uses population-based longitudinal designs to understand the determinants and consequences of suicidal behavior and other mental health problems across the lifecourse. This talk will focus on her recent work using statewide emergency department data from California to examine health, healthcare utilization, and mortality outcomes among adolescents and adults who self-harm. The implications of her findings for treatment models and suicide prevention programs in emergency care settings will be discussed.  


yiNengjun Yi, PhD | View Seminar
Professor, Department of Biostatistics
Sir David Cox Endowed Professor in Biostatistics
University of Alabama, Birmingham

“Hierarchical Models for Microbiome Data Analysis”

Microbiome/metagenomic data generated by the next-generation sequencing (NGS) technology provide valuable resource for investigating associations between the microbiota and host clinical/environmental variables. However, the development of statistical methods and computational tools for properly analyzing and interpreting these data is in its infancy. In addition to the well-known properties (for example, compositional count data structure, over-dispersion, and zero-inflation), microbiome studies usually measure many correlated host variables, and collect samples with spatial and temporal dependences (for example, longitudinal studies). These properties have important implications in the analysis and interpretation of microbiome data, but have not been fully addressed with efficient methods.
We have recently developed hierarchical negative binomial and zero-inflated negative binomial models to jointly analyze many correlated host clinical/environmental variables. We also have developed negative binomial mixed models and zero-inflated negative binomial mixed models for analyzing microbiome data with spatial and temporal dependences. We have implemented these methods in our R packages, BhGLM, bzims, and NBZIMM. I describe these models and R packages, and demonstrate their applications to real data.


thompsonMike Thompson, PhD, MPH | View Seminar
Assistant Professor, Cardiac Surgery
University of Michigan

"Applying Epidemiologic Methods to Improve Value in Cardiac Surgery"

Nearly 300,000 Americans undergo cardiac surgery every year, and it remains among the most hazardous and expensive episodes of care. As health care payers continue to push value-based reimbursement in cardiac surgery, it is imperative that we identify opportunities to reduce excessive health care spending without sacrificing quality of care. By applying epidemiologic principles, we can identify these opportunities across patients and providers through quantitative analysis of clinical and administrative data. In this talk, I will highlight some of these newly discovered opportunities, and how hospital collaborative learning is being leveraged to improve the value of cardiac surgery in Michigan.