ITC displacement titrations and direct ITC titrations at low c-value can be used as a reliable technique to study low-affinity interactions as usually given for the binding of a fragment to a target protein. Important aspects can be concluded from this study. First of all, for ITC displacement titrations the fragment must bind to a part of the protein binding pocket that overlaps with the binding pose region of the applied reference ligand. Secondly, the selected reference ligand must have a significant higher or lower binding enthalpy so that a heat difference signal can be recorded for the fragment. Furthermore, it has to be regarded that, as for any displacement titration, all errors affecting the parameter determinations of the reference ligand will add to the accuracy of the parameters obtained for the fragment. Finally, if no suitable reference ligand is available and the fragment binds enthalpically enough, low c-value titrations can be applied as an alternative. In this case, however, some anticipated knowledge about the expected binding affinity of the fragment must be available to estimate the required excess concentration of the fragment at the end of the titration. The solubility of the fragment and the protein are the most crucial issues in low c-value titrations because the required high concentrations of the fragment to be studied in the injection syringe should exceed the final sample cell concentrations with respect to the fragment’s KD value. In order to estimate the required concentration of the fragment, a crude estimate of the fragment’s KD obtained from an independent experiment is highly recommended. It is very important that measured binding enthalpies should not be compared quantitatively across different experimental conditions, but only relative to each other by using one suitable measurement protocol. Remarkably, different reference ligands used for the displacement of pre-incubated fragments reveal deviating enthalpic signals. Most likely these differences are caused by differences in the water structure and thus in the residual solvation pattern at the binding site of the different ligands used for titration. The consideration of the relative enthalpy differences across the different displacement experiments showed a consistent relative difference. This allows to characterize the studied fragments relative to each other. The detailed interpretation of the enthalpic signature requires a comparison of the corresponding crystal structures in order to select fragments as superior enthalpy-dominated candidates for further development. To achieve this goal, high-resolution crystal structures of six fragments binding to the S1 pocket of thrombin were determined and analyzed with respect to the thermodynamic binding profiles. Binding affinity is not only determined by the properties of the formed protein-ligand complex but some differences can already discriminate ligands in aqueous solution prior to any protein binding. Conformationally restricted ligands can experience a significant binding advantage over more flexible ligands if the correct protein-bound geometry is well preorganized in a rigidified skeleton. A new derivative of a 3-amidinophenylalanine matriptase inhibitor was crystallized with thrombin and the resulting complex was superimposed with a representative X-ray crystal structure of matriptase bound to a related inhibitor. From the superimposition of these structures information could be extracted regarding the selectivity profiles towards serine proteases observed in this study. Due to the large homology of amino acids in the binding pocket of thrombin and matriptase, reliable predictions about the possible binding mode of the inhibitor, can be made for matriptase.