Computing artificial backbones of hydrogen atoms in order to discover protein backbones

NMR experiments are able to provide some of the distances between pairs of hydrogen atoms in molecular conformations. The problem of finding the coordinates of such atoms is known as the molecular distance geometry problem. This problem can be reformulated as a combinatorial optimization problem and efficiently solved by an exact algorithm. To this purpose, we show how an artificial backbone of hydrogens can be generated that satisfies some assumptions needed for having the combinatorial reformulation. Computational experiments show that the combinatorial approach to this problem is very promising.