The search algorithm is based on the calculation of similarity metrics between the query structure and the database structures represented as MNA and QNA descriptors, respectively. MNA descriptors represent the structure as a set of atom-centered fragments. QNA descriptors are based on electron affinities and ionization potentials of isolated atoms as well as on the connectivity matrix of the molecule. QNA descriptors characterize each atom in a molecule, taking into account the influence of all other atoms of the molecule on this atom. The algorithm for calculating similarity between structures represented by MNA descriptors uses the Tanimoto coefficient, and for structures represented by QNA descriptors, the Todeschini approach is used.