In Gram-negative bacteria, outer-membrane protein channels, such as OmpF of Escherichia coli, constitute the entry point of various classes of antibiotics. While antibacterial research and development is declining, bacterial resistance to antibiotics is rising and there is an emergency call for a new way to develop potent antibacterial agents and to bring them to the market faster and at reduced cost. An emerging strategy is to follow a bottom-up approach based on microscopically founded computational based screening, however such strategy needs better-tuned methods. Here we propose to use molecular dynamics (MD) simulations combined with the metadynamics algorithm, to study antibiotic translocation through OmpF at a molecular scale. This recently designed algorithm overcomes the time scale problem of classical MD by accelerating some reaction coordinates. It is expected that the initial assumption of the reaction coordinates is a key determinant for the efficiency and accuracy of the simulations. Previous studies using different computational schemes for a similar process only used one reaction coordinate, which is the directionality. Here we go further and see how it is possible to include more informative reaction coordinates, accounting explicitly for: (i) the antibiotic flexibility and (ii) interactions with the channel. As model systems, we select two compounds covering the main classes of antibiotics, ampicillin and moxifloxacine. We decipher the molecular mechanism of translocation of each antibiotic and highlight the important parameters that should be taken into account for improving further simulations. This will benefit the screening and design for antibiotics with better permeation properties.