Modern cosmology successfully deals with the origin and the evolution of the Universe at large scales, but it is unable to completely answer the question about the nature of the fundamental objects that it is describing. As a matter of fact, about 95% of the constituents of the Universe is indeed completely unknown: it cannot be described in terms of known particles. Despite intense efforts to shed light on this "literal" darkness by dark matter and dark energy direct and indirect searches, not much progress has been made so far. In this work, we take a different perspective by reviewing and elaborating an old idea of studying the mass-radius distribution of structures in the Universe in relationship with the fundamental forces acting on them. As we will describe in detail, the distribution of the observed structures in the Universe is not completely random, but it reflects the intimate features of the involved particles and the nature of the fundamental interactions at play. The observed structures cluster in restricted regions of the massradius diagram linked to known particles, with the remarkable exception of very large structures that seem to be linked to an unknown particle in the sub-eV mass range. We conjecture that this new particle is a self-interacting dark matter candidate.