Conserved Secondary Structures in Viral mRNAs

Figures

• 

  Figure 1. Venn diagram showing the taxonomy of the host organisms within all viral orthologous groups (VOGs). Only those VOGs are included for which host annotation for all viruses is available in the Virus-Host DB.

• 

  Figure 2. Distribution of VOG sizes.

• 

  Figure 3. Overview of the analysis of conserved RNA structures in VOGs.

• 

  Figure 4. Number of local RNA structures as a function of VOG size.

• 

  Figure 5. Coverage of VOG alignments by local optimal RNA structures. (a) VOGs with two sequences. (b) VOGs with more than two sequences, in which structures are conserved across all sequences. (c) subVOGs. VOGs that did not contain conserved structures, even after splitting into subVOGs, are not shown.

• 

  Figure 6. Taxonomic distribution of proteins in VOGs (with more than two sequences) and subVOGs.

• 

  Figure 7. Taxonomic distribution of hosts in VOGs (with more than two sequences) and subVOGs.

• 

  Figure 8. Structural coverage as a function of the taxonomic variety of subVOGs and their host organisms.

• 

  Figure 9. Example of a VOG split into structurally homogenous subVOGs. Shown is the VOG 00052 containing 20 mRNAs, encoding for Kila-N domain proteins, from 12 virus species. On the left, the neighbor-joining tree based on the pairwise sequence identity between the protein sequences is shown. Colored boxes indicate subVOGs, within which conserved structures were predicted. The tree nodes outside colored boxes did not yield any conserved structures. On the right, the structure conservation index (SCI) (black line for each subVOG alignment) is plotted against the alignment position on the percentage scale. Plots are ordered according to the subVOG position in the tree.

• 

  Figure 10. Structures found in Influenza A and B mRNAs encoding the matrix protein (VOG11160). Colors in MSA pictures encode compensatory mutations supporting the consensus structure. Red marks pairs with no sequence variation; ochre, green, turquoise, blue, and violet mark pairs with 2, 3, 4, 5, and 6 different types of pairs, respectively. (a) The second of the two consecutive stem loops of the structure proposed by Jiang et al. [52], covering positions 147–192, visualized with R2R [54]; (b) The predicted conserved consensus structure for nucleotides 148–188 supports the second hairpin loop of the model of Jiang et al., shown in (a). Colors encode the positional entropy; (c) Structure-guided alignment and dot bracket structure notation for the consensus structure shown in (a). The upper sequence corresponds to Influenza A and the lower sequence to Influenza B; (d) Shown are two consecutive hairpin loops for nucleotide positions 682 to 744, proposed by Moss et al. [23], visualized with R2R; (e) The predicted conserved structure for nucleotides 697–758 partly supports the model shown in (e). Colors encode the positional entropy; (f) Structure-guided alignment and dot bracket notation for…

• 

  Figure 11. Example structures that were identified within subVOGs. (a) Structural annotation of the subVOG 30, belonging to VOG00029, which contains six mRNAs encoding a replicase protein of different Tobamovirus species. Consensus structure visualized by RNAalifold. Colors encode the positional entropy; (b) Structure-guided MSA and consensus structure in dot bracket notation corresponding to consensus structure shown in (a). Colors encode compensatory mutations supporting the consensus structure. Red marks pairs with no sequence variation; ochre, green, turquoise, blue, and violet mark pairs with 2, 3, 4, 5, and 6 different types of pairs, respectively; (c) Consensus structure of subVOG 64 from VOG00003, which contains four mRNAs coding for a p28-like protein of different alphabaculoviruses; (d) Structure found in a Heliothis virescens ascovirus 3e, by covariance model search of the structure shown in (c), using cmsearch in the entire sequence space of all VOGs.

• 

  Figure 12. MRNA folding energy as a function of (a) sequence length and (b) GC-content. DGmin: Minimum folding energy of either all possible 30-nucleotide windows of a sequence or all found local optimal structures using RNALfold. DGmean and DGmax: Mean and maximum of all windows, respectively.