GOLD fails with the error message "do_ullman: query size larger than target"

The "do_ullman" error message refers to a substructure matching gone wrong.

This error message is obtained when a reference ligand (i.e. usually the crystallographically observed binding pose) is specified in order to calculate docking RMSDs for the different docking poses to this  reference ligand. In the gold.conf file, this is indicated on a line such as:

ligand_reference_file = <filename>

GOLD is not able to recover from a fatal error, in that it is not able to skip to the next ligand. Strictly speaking, it is not possible to restart a GOLD job from the position it stopped.

However, if you are docking multiple ligands from a single multi-sd file, then it is possible to start a new GOLD job that begins docking ligands from a user specified position in the multi-sd file. You would need to edit the ligand_data_file line in your gold.conf file and specify the ligand you wish to start docking at. For example, if you were docking 100 ligands contained in a multi-sd file (with 10 GA runs per ligand) but GOLD encountered an error such that only the first 22 ligands were docked fully and the 23rd ligand did not finish docking, you can submit a new GOLD run by editing your gold.conf file such that the ligand_data_file line would follow such an example:

ligand_data_file C:/path/my_100_ligands.sdf 10 start_at_ligand 23