With the swift advance in exoplanet sciences it is now possible to characterize not only the fundamental parameters (mass and radius) of planets but also their interior structure and bulk composition. The former is known to influence on the habitability conditions of terrestrial planets, and the latter in itself is a key aspect to understand planet formation processes and the origin of their diversity.In order to accurately assess planetary internal composition, the derivation of the chemical abundances of the host stars is of extreme importance. In particualr, stellar abundances of Fe, Si, Mg are proposed as principal constraints to reduce degeneracy in exoplanet interior structure models under assumption of identical composition of these elements in the rocky planets and their host stars.This regularly used assumption is based on our knowledge that stars and planets form from the same primordial gas and dust cloud. It is also supported by our Solar System observations for which we know that the composition of major rock forming elements (such as Mg, Si, and Fe) in the meteorites and telluric planets (with the exception of Mercury) is similar to that of the Sun. However, direct observational evidence for the aforementioned assumption for exoplanets is absent.By using the largest possible sample of precisely characterized low-mass planets and their host stars, we show that the composition of the planet building blocks indeed correlates with the properties of the rocky planets (see Fig. 1). We also find that on average the iron-mass fraction of planets is higher than that of the primordial values, owing to the disk-chemistry and planet formation processes. Additionally, we show that super-Earths and super-Mercuries appear to be distinct populations with differing compositions, implying differences in their formation processes. We suggest that giant impact alone is not responsible for the high-densities of super-Mercuries.I propose an oral contribution to speak about these very recent results published in Scinece.Fig. 1 The iron-mass fraction of the planets inferred from the planets' mass and radius as a function of the iron-mass fraction of the protoplanetary disk, estimated from the host star abundances. Super-Mercuries (in brown) and super-Earths (in blue) appear as two distinct groups.