Modelling is essential in the engineering of digital twins: A digital twin includes models and data representing the real-world entity (`models in digital twins`) and models describing and operating the digital twin as a software system (`models of digital twins`). Consequently, we need to systematically develop and manage these different kinds of models over time. Model-driven Engineering (MDE) offers methods to address these challenges. Here, we provide an overview of the article 'Model-Driven Engineering for Digital Twins: Opportunities and Challenges', published in the INCOSE Systems Engineering journal. It summarizes outcomes from the Dagstuhl Seminar 22362 'Model-Driven Engineering of Digital Twins' and provides insights into the digital twin life cycle, the use and re-use of different model kinds across life cycle stages, and a research roadmap covering six key aspects of digital twin research using MDE.