obfit question

Hi,
    Obfit tool performs better than the LigAlign plugin I used before. Yet, the smarts pattern has to be supplied to be a matching region. I test massive cases which the hits were superimposed to the ligand conformation in crystal structures. Smile formats were generated from ligands to be input as the smarts expression. Resultantly,  no translation were operated. The culprit I guess is the different charge information between the pivot and querys. Is there an easy way to handle the tiny problem? one typical case is attached.