Characterizing origins of drug resistance will ultimately enable the design and optimization of improved compounds. We are using computational simulation methods, followed by energetic and structural analysis, to probe the binding interfaces of protein-ligand complexes to determine which features are most important for molecular recognition and/or change upon point mutation. All-atom molecular dynamics, MM-GBSA binding energy estimation, and per-residue molecular footprinting are used to delineate which residues loose Coulombic energies, van der Waals energies, or H-bonds interactions at specific sites. Results will be presented for small molecule inhibitors which target the viral enzyme neuraminidase, the antic-cancer target EGFR, and for peptide inhibitors of the HIV glycoprotein gp41. Progress towards development of a docking test-set aimed at improving virtual screening will also be presented.