Pleiotropy: Epidemiologic analyses and implications for return of results

Abstract

Pleiotropy refers to a single gene with multiple phenotypic effects. Genetic variants associated with multiple diseases can point to shared biologic mechanisms or other similarities. The first half of this dissertation comprises two epidemiologic analyses evaluating common genetic variants for pleiotropic associations. The first investigates whether levels of the inflammatory biomarker C-reactive protein are associated with genetic variants previously associated with various inflammation-related diseases such as stroke, obesity, or type 2 diabetes. The second evaluates whether genetic variants previously associated with various cancers are also associated with melanoma. Results from both analyses provide evidence supporting both known and potentially novel pleiotropic associations. The second half of this dissertation focuses on how the presence of pleiotropy impacts the return of individual genetic results to research participants. Researchers are generating increasingly large amounts of genetic information on research participants, uncovering genetic variants with a spectrum of health implications. While many genetic variants are of little or unknown clinical significance, some genetic variants are highly predictive of disease, and can inform screening or other actions that can prevent or reduce disease. A few variants are sufficiently important to suggest an ethical duty on the part of the researcher to return that genetic information to a participant. Conversely, other variants may impart information on risk that is not actionable, and returning such information may inflict more harms than benefits. Recent guidelines have proposed various criteria for prioritizing which results are appropriate or not appropriate for return, focusing on weighing the clinical validity and clinical utility of a particular genetic result. However, these guidelines do not currently take into account that genetic variants may have multiple meanings, which might have contradicting recommendations for whether that same variant should be returned. This dissertation reviews the concept and extent of pleiotropy, reviews current recommendations for returning results, and explores the ways in which the existence of pleiotropic variants impacts these guidelines. This dissertation proposes a potential framework for considering pleiotropic relationships in return recommendations, and suggests that future guidelines will need to account for pleiotropy if they are to be effective.