I have a series of irreversible covalent kinase inhibitors. Data shows the covalent interaction of inhibitor with active site residue play crucial role in activity value. Figure shows the most active compound covalently bonded to the CYS909 residue.

Is it reasonable to develop 3D-QSAR model based on this data. Given that 3D-QSAR techniques relates activity data to change in steric and elctrostatic fields, hydrophobicity etc. but not covalent bonding specifically.

Kindly give your arguments !

Thank you