NMR spectroscopy is a technique which exploits magnetic properties of atoms and molecules to determine their physical and chemical properties. In this particular case, the NMR spectroscopy is a part of a system for determining structure of proteins, where the latter finds many applications in medical research, drug design etc. NMR spectroscopy produces large amounts of data that need to be automatically processed. One of the main problems is mapping of chemical shifts (peaks in NMR spectra) to particular atoms in a given protein. The available data set is rather complex and the assignment problem non-trivial. The aim is to develop (implement and test) methods, possibly based on existing machine learning techniques, for solving the assignment problem.