Biotechnological hydrogen production by enzymes is a promising alternative energy source, either by direct generation of biogas or coupled to fuel cells. Nevertheless, most of the hydrogenase enzymes are easily inactivated by traces of O-2, generating spectroscopically detectable non-catalytic species, diffusible reactive oxygen species, or directly by oxidation of the protein matrix. To implement technologies based on hydrogenases, the atomic and electronic features that rule the electron transfer process still need to be identified and understood. We studied the electron transfer pathways between the metal cluster and protein matrix of [NiFe]-hydrogenase from Salmonella enterica using multi-level molecular simulation: all-atom molecular dynamics, quantum mechanics/molecular mechanics, and with bioinformatics tools, such as statistical coupled analysis, to identify key residues of the process. The identification of these residues could be a more advantageous alternative compared with random point mutations, in order to improve catalytic performance or operational stability by genetic engineering. (C) 2017 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.