Qian-Rui Huang showed me this remarkable beaded model for the (4,3) torus knot today. The model is not like the one for trefoil knot. It is quite soft and very crowded in the center since two tubes have to pass to each hole for the (4,3) torus knot. This problem is not very serious for the trefoil knot though. As I said before, one can in principle construct any kind of molecular graph. But only in certain situations, the resulting structure can simulate the real fullerene.