What is Tortoize?
Tortoize checks the torsion angles in your protein model and gives them a Z-score based on their distribution in high quality models in the PDB-REDO databank. Higher Z-scores are better. For each residue, two sets or torsion angles are checked: φ/ψ combinations (the Ramachandran plot) and χ1/χ2 combinations(the rotamers). The distributions that are use to check torsion angles are residue en secondary structure (as assigned by DSSP) specific. The strength of Tortoize is that assigns global, per-model Z-scores on top of the per-residue scores (the Ramachandran Z-score, or Rama-Z and the rotamer Z-score). These global scores tell you how normal your Ramachandran plot and your rotamer distribution are compared to high-quality structure models. And, unlike per-residue scores, these global scores are not sensitive to over-enthusiastic use of restraints in model making.
The Tortoize papers and code
| Tortoize paper | https://www.cell.com/structure/fulltext/S0969-2126(20)30287-2 |
| Original algorithm | https://academic.oup.com/bioinformatics/article/13/4/425/274735 |
| Our initial implementation | https://journals.iucr.org/m/issues/2018/05/00/jt5027/ |
| The Tortoize code | https://github.com/PDB-REDO/tortoize |