NovaFold report

When you open a completed NovaFold prediction (.novafold) in Protean 3D, a specialized Report view contains detailed results about the prediction.

The sections of the NovaFold Report are:

• Molecule
• Templates
• Model Overview
• Model ‘x’ and its subsections:
  • Similar Experimental Structures
  • Predicted Binding Sites
  • Predicted Enzyme Function
  • Predicted Protein Function

The following video is a brief introduction to the NovaFold Report:

When viewing prediction results in the Report view, it may seem natural to compare the similarities between each section of a model and a template. However, this is most appropriate for molecules where a homolog (both similar in protein fold and sequence) is known. NovaFold uses a “template ensemble” approach, where different pieces of multiple templates are assembled in different combinations. This improves the structural diversity of the sampled structural conformations for the query molecule, which leads to more accurate models for molecules without clear homologous templates.

In addition, NovaFold also uses a model refinement technique when building the final all-atom models. By design, it modifies and repacks local regions in the model looking for lower-energy conformational intermediates.

In summary, there are multiple algorithms in play which are designed to distort the final models away from the starting templates in search for the lowest-energy conformations.

To copy text from the NovaFold Report, see Use basic editing commands.

To open a template or structure from the report, see Open a structure from the NovaFold Report.