Fall 2021 Seminars

In the United States (US), the prevalence of recent cannabis use has risen steadily over the past decade and coincides with the growing acceptability of and access to cannabis across the country. During this time, states with approved medical cannabis laws became the majority. Notably, Veteran groups were among those who advocated for such state legislation, and major qualifying conditions for many state medical cannabis programs are common among the US Veteran population (including pain and posttraumatic stress disorder [PTSD]). Nonetheless, compared with the US general population, relatively little is known about cannabis use and health among Veterans. In this seminar, Dr. Bohnert will present findings from recent and ongoing work on non-medical and medical cannabis use and health issues among Veterans and nonveterans in the US. Research to be discussed includes findings from large, nationally representative surveys, as well as results from a longitudinal cohort study examining patterns of cannabis withdrawal symptoms. Dr. Bohnert’s research in this area is supported in part by a Department of Veterans Affairs (VA) Health Services Research and Development Service (HSR&D) Investigator Initiated Research (IIR)/Merit Award.

Despite the great progress achieved through Genome-Wide Association (GWA) studies most risk-loci remain unmapped. Single marker regression (SMR), the method most often used in GWA studies, tests the association between SNPs and phenotypes one-SNP-at-a-time. Achieving high power with SMR requires using a large sample size and tight linkage disequilibrium (LD) with causal variants. Unfortunately, these conditions also reduce the mapping resolution of SMR making this method poorly equipped for fine mapping of risk variants with big data. Bayesian variable selection (BVS) methods can offer high power with precise mapping resolution and accurate control of the false-discovery rate (FDR). In this study we evaluate the power-FDR performance of a BVS procedure for mapping very-small-effect risk variants with biobank-size data. Using simulations, we show that with large sample size the BVS regression can achieve high power with low FDR (a performance comparable to that of Forward regression) and a much better mapping resolution than SMR and LASSO. We also show that with biobank-size data, SMR can lead to very poor mapping resolution and a very high FDR. We demonstrate the use of BVS for fine mapping with biobank-sized data by applying the methods and software presented in this study to map risk loci for six blood-biomarkers using data from the UK-Biobank (~300,000 samples and ~5.5 million SNPs). The BVS procedure identified a larger number of chromosome segments with considerably finer mapping resolution than SMR. The study is accompanied with R-software and data visualization tools that can be used for GWA analyses with biobank-size data.

 

The effect of racial group assignment on preterm delivery is robust. The question is ‘how’– the causal mechanism. Socioeconomic position (SEP) affects experiences of risk and protective factors throughout the lifecourse; but would a change in SEP change Black-White disparities? The objective of this study is to quantify the potential contribution, on racial disparities in preterm delivery, of an intervention (Child Development Accounts, aka BabyBonds) that would change the distribution of family wealth over the lifecourse.

The under-enrollment of women in randomized controlled trials (RCTs) has been a long-standing issue in clinical medicine, particularly in cardiovascular disease where sex-based disparities still persist. In the past decade, clinical trials have been the origin of major advances in the treatment of acute stroke, especially mechanical thrombectomy. However, given the well-documented importance of sex in the epidemiology and pathophysiology of stroke, the possible under-enrollment of women in these and other stroke trials represents a threat to their validity and generalizability. In this presentation I will give an overview of the results of a meta-analysis examining sex differences in the enrollment of women in acute stroke trials conducted over the last decade. I will quantify the magnitude and certainty of the observed differences, illustrate data that points to the potential origins, and discuss the broader literature on the enrollment of women in trials and how this could be improved upon.