InterPro entry page

The InterPro entry page consists of the following features (see figure 5):

• Entry type and name
• Contributing signatures
• Overlapping homologous superfamilies
• Hierarchical relationships to other InterPro entries
• Description of the entry, with links to references
• GO terms that have been associated with that entry. GO terms are divided in three categories: biological process, molecular function and cellular component.

Using the tabs below the entry name, information can be found about proteins matched by that entry, their domain organisation, their taxonomic coverage (i.e. the species in which the proteins are found), which proteomes are represented, the pathways and interactions proteins matched by that entry are involved in, any known experimental (Structures tab) or predicted structures (AlphaFold tab) for those proteins, and also any predicted models from the Genome3D consortium.

Exploring the different tabs in the InterPro in entry page for the Dynamin-2 family (IPR027188) in the InterPro website, answer the following questions.

Questions

Experimental structures

The Structures tab gives the list of experimental structures matching the InterPro entry.

For each structure, an indication of the region of the entry matched by the structure is provided in the MATCHES column, hovering over a region displays more information (e.g. start/end position). The Structure entry page can be accessed by clicking on the structure identifier, name or the miniature image of the structure.

Looking at the Structures tab of the Dynamin-2 family, answer the following questions.

Questions

Predicted Structures

The AlphaFold tab provides predicted structures for proteins matching the InterPro entry generated by DeepMind. The advantage of predicted structures is that they are generated for the full length of the protein, while experimental structures don’t necessarily map the whole length, given us a better idea of the different domains making the protein.

The prediction should be looked at in consideration with the confidence scale provided, as some regions of the protein can be more difficult to model.

As you may be aware, domains with similar 3D structures can have similar function. Based on this assumption, each domain in the predicted structure can potentially be compared to experimental structures in PDBe (e.g. using PDBeFold) to hypothesise a function.

By default one of the prediction is displayed in the 3D viewer, the full list of predictions is available at the bottom of the page.