Even a single biological macromolecule is a complex object. In many cases it is neither possible nor desirable to describe the dynamics of intra- and intermolecular interactions on biologically relevant timescales with ab initio approaches. In these situations, statistical physics offers a rich set of useful tools to quantitate biological processes using coarse-grained models. These models can incorporate the "mechanical" properties of molecular biological systems as well as the biological information encoded in sequences or specific interactions. I will discuss the interplay of these two aspects in two exemplary problems.