This study has taken a closer look at the theoretical basis for protein-fragment interactions. The approach involved the deconstruction of 3 non-nucleoside inhibitors of HIV-1 reverse transcriptase and investigation of the interaction between 21 substructures and the enzyme. It focused on the concept of ligand efficiency and showed that ligand independent free energy fees (ΔG(ind)) are crucial for the understanding of the binding affinities of fragments. A value of 7.0 kcal mol(-1) for the ΔG(ind) term is shown to be a lower limit for the NNRTI binding pocket of HIV-1 RT. The addition of the ΔG(ind) term to the dissociation free energy in the calculation of a corrected ligand efficiency, in combination with the lack of an efficient ligand binding hot spot in the NNIBP, fully explains the existence of nonbinding NNRTI substructures. By applying the concept to a larger set of ligands, we could define a binding site profile that indicates the absence of an efficient fragment binding hot spot but an efficient binding of full-sized NNRTIs. The analysis explains some of the challenges in identifying fragments against flexible targets involving conformational changes and how fragments may be prioritized.