Common operational issues (and fixes)

“My run seems stuck”

• Check Status with the Job ID.

• Large inputs + large cutoff can make extraction/training slow; try a smaller test first.

“Outputs are empty or extremely small”

• Verify the structure actually contains RNA residues and the selected atom (e.g., C3').

• Reduce sequence separation (too large can remove many contacts).

• Ensure chains are valid (in paste mode, wrong chain IDs often yield empty extraction).

“Scores look inconsistent across runs”

• Ensure you did not change cutoff/bin width/atom mode/separation.

• Ensure the training set is the same (different training structures = different model).

“I want to score many test sets with the same model”

• Run Extract + Train once (or a Pipeline run).

• Then repeatedly use the Score tab with new structures, keeping parameters identical.