Integrating 3D Structures into a SEED Alignment

Updating an Rfam family to include 3D structures involves several steps. This document outlines the process.

Acronyms

  • ncRNA: non-coding RNA

  • PDB: Protein Data Bank

  • SS: Secondary Structure

  • SS_cons: Alignment secondary structure consensus

  • RF: Alignment reference sequence

  • PK: Pseudoknot

Validation

Validate that the 3D information has the correct identifiers. The alignments map the PDB sequences to RNAcentral IDs. The format for PDB sequence reference includes RNAcentral ID plus coordinates. The PDB secondary structure reference includes the sequence reference plus the PDB reference.

For example:

URS0000A76344_911092/1-94
#=GR  URS000080DF35_32630/1-94      2GIS_A_SS

Where URS0000A76344_911092 is the RNAcentral reference: https://rnacentral.org/rna/URS0000A76344/911092 and 2GIS_A is the PDB reference: https://www.rcsb.org/structure/2GIS

Review and Update of 3D References

  1. Validate that PDB structures have high resolution (<4.0Å). If the added structures are of low resolution or are missing a secondary structure, remove the sequence.

  2. Review and update base pairs in the sequence. Not all base pairs may be annotated correctly. For example, modified nucleotides are not annotated in the 2D but may be valid base pairs. Base pairs between modified nucleotides are included manually in the PDB SS annotations.

  3. Review and update regions involved in contact with other chains, RNA, or proteins. In structures that report an interaction between two RNA chains, the PDB 2D may include incorrect base pairs. The base pairs will be between the two chains but incorrectly assigned to one of them. These must be removed. In the case of interactions with proteins, this interaction can prevent base pairs in the PDB SS; these regions have to reflect the 3D information in the PDB SS but may not be considered for the SS_cons.

  4. Review ‘canonical’ base pairs. The PDB 2D must include all G:C, A:U, and G:U base pairs reported in the structure. If a base pair is missed, it is added; if a base pair is not supported, it is removed.

  5. Review PKs. The PKs are represented in the PDB SS as open {{{ and closed }}} brackets. This has to be confirmed in the structures and updated if needed.

Update the SS_cons Line

The SS_cons attempts to represent what is reported in the structures and is updated after reviewing each of the PDB SS. The updates can reflect:

  • PKs annotated in the PDB SS; PK base pairs need to be included considering the base pairs supported in the alignment and the RF sequence.

  • Single base pairs that are not stem loops but are supported by the structures. These base pairs are less easily caught by classical alignment methods, but recognizing them in the 3D structures and adding them to the SS_cons helps direct the alignment and represent a more accurate structure.

  • Missing base pairs in stem-loops, for example, stem-loops with variable length. Whenever the SS_cons can represent as many base pairs in a variable length of the stem-loops, this has to be included to provide the best-supported structure.

  • Base pairs in riboswitches or RNAs with ligands. Some PDBs with ligands prevent the formation of base pairs in their bound conformation. If there is support for base pairs in an unbound structure, these base pairs have to be included in the SS_cons.

  • When the RNA is compromised by interacting with proteins or RNAs, the SS_cons needs to be corrected to avoid reflecting extra structural elements that are not supported by the structures.

Refine the Alignment

As a result of the curation process, the length of the sequences may need to be adjusted, and the alignment may need to be refined to better align variable regions. Some supported refinements are:

  1. When the 3D structure supports a longer or shorter ncRNA, consider extending or trimming the alignment to represent a more accurate model. The sequence length has to be reflected in all the alignment sequences, and the improvement in the family has to be confirmed in terms of keeping or improving the annotation of the family in the current annotation dataset of Rfam.

  2. Confirm that only canonical Watson-Crick base pairs are represented in the RF. In some cases, it may be necessary to refine the alignment. This is done with Infernal, providing the reviewed SS_cons as a reference.

  3. If the 3D structures report new elements, for example, PKs or missed base pairs that are not represented in the RF, the alignment has to be refined to correctly represent these elements. This is done with Infernal, providing the reviewed SS_cons as a reference.

Add Structural Annotations

Annotations are not mandatory in the alignment; they are provided as a guide to users when the structure has been defined in terms of 3D structural elements (#=GC RNA_structural_elements). Ligands and RNA motifs can also be annotated as ‘x’ in the positions that form the motif (example: K-turn) or the position that represents contacts of the sequence with the ligand (#=GC RNA_motif and #=GC RNA_ligand). Since authors use several nomenclatures for structural elements, the structural elements have a free format. Some examples are: stem loop, SL, stem, S, P, etc.

Update the DESC File

The literature reference of each structure in the SEED has to be added to the family DESC, and the PDB references have to be included in the DESC CC lines.

Basic References

  • Canonical base pairs: A:U, C:G, and G:U are considered canonical base pairs.

  • Non-canonical base pairs: Non-canonical base pairs are kept in the PDB SS as a reference of the PDB reported sequence but are not in the SS_cons reference.

  • Pseudoknots: PKs in PDB SS are represented by {{{ and }}}, PKs in SS_cons are represented by AAA aaa, BBB bbb, CCC ccc, etc.

Whenever annotations of motifs, ligands, or structural elements can be included, they can be annotated with the following #=GC lines:

#=GC RNA_motif
#=GC RNA_ligand
#=GC RNA_structural_elements