"The moduli space of translation surfaces in fixed genus is an orbifold endowed with a SL(2,R)-action preserving a probability measure. It was shown by Masur and Veech that the this action is ergodic on each connected component of the moduli space. As an analogue of Ratner's theorem, Eskin and Mirzakhani proved a structural result for any SL(2,R)-invariant measures and orbit closures. More precisely, they show that any SL(2,R)-orbit closure is an orbifold that supports a unique SL(2,R)-invariant probability measure. However, contrarily to Ratner's theorem, their result does not give a recipe to compute the list of all SL(2,R)-orbit closures. The construction of SL(2,R)-invariant orbifolds in the moduli space of translation surfaces is a very active line of research. In a joint work with J. Rüth and A. Wright we build a new example of such orbit closure in genus 8 which we believe is the last exceptionnal example coming from quadrilateral unfolding. In this talk I will review Eskin-Mirzakhani result in parallel to Ratner theorem, quickly mention one motivation for understanding SL(2,R)-orbit closures (dynamics of rational billiards) and finally explain our construction."