When I started grad school in 2003 there were a handful of structures approaching 0.4 nm, which is the point where you might be able to build most of the sequence into the map de novo. But those were exceptional cases.