A Reference Meta-model to Understand DNA Variant Interpretation Guidelines

Abstract

Determining the role of a DNA variant in patients' health status -a process known as variant interpretation -is highly critical for precision medicine applications. Variant interpretation involves a complex process where, regrettably, there is still debate on how to combine and weigh diverse available evidence to achieve proper and consistent answers. Indeed, at the time of writing, 22 different variant interpretation guidelines are available to the scientific community, each of them attempting to establish a framework for standardizing the interpretation process. However, these guidelines are qualitative and vague by nature, which hinders their streamlined application and potential automation. Consequently, more precise definitions are needed. Conceptual modeling provides the means to bring clarification within this domain. This paper presents our efforts to define and use a UML meta-model that describes the main concepts involved in the definition of variant interpretation guidelines and the constructs they evaluate. The precise conceptual definition of the guidelines allowed us to identify four common misinterpretation patterns that hamper the correct interpretation process and that can consequently affect classification results. In several proposed examples, the use of the meta-model provides support in identifying the inconsistencies in the observed process; this result paves the way for further proposing reconciliation strategies for the existing guidelines.