Table 1.
Attainment of native-like structure by de novo Fragment Assembly of RNA with Full Atom Refinement (FARFAR), using the full-atom Rosetta energy function. The lowest energy 500 of 50,000 refined conformations were clustered with a model-model heavy-atom RMSD cutoff of 2.0 Å. The five lowest energy clusters were taken as the de novo models; features of the best cluster (lowest RMSD to the experimental structure) are listed. See Supplementary Fig. 2 for motif definitions.
| Motif properties |
Clustering statistics |
Cluster center | Lowest energy cluster member |
Lowest RMSD sampled |
|||||
|---|---|---|---|---|---|---|---|---|---|
| No. res. |
No. chains |
Clust Rank |
Cluster size |
RMSDa | fNWCb | RMSDa | fNWCb | ||
| G-A base pair | 6 | 2 | 1 | 471 | 1.19 | 1/ 1 | 1.89 | 0/ 1 | 0.54 |
| UUCG tetraloop | 6 | 1 | 1 | 498 | 1.12 | 1/ 1 | 1.14 | 1/ 1 | 0.64 |
| GAGA tetraloop from sarcin/ricin loop | 6 | 1 | 1 | 500 | 0.82 | 1/ 1 | 1.00 | 1/ 1 | 0.52 |
| Loop 8, A-type Ribonuclease P | 7 | 1 | 5 | 27 | 1.38 | 0/ 0 | 1.41 | 0/ 0 | 1.13 |
| Pentaloop from conserved region of SARS genome |
7 | 1 | 3 | 237 | 1.10 | 1/ 1 | 1.48 | 1/ 1 | 0.88 |
| L3, thiamine pyrophosphate riboswitch | 7 | 1 | 4 | 6 | 2.00 | 0/ 1 | 2.68 | 0/ 1 | 1.44 |
| Fragment with A-C pairs, SRP helix VI | 8 | 2 | 1 | 284 | 1.83 | 2/ 2 | 2.74 | 1/ 2 | 0.48 |
| Helix with U-C base pairs | 8 | 2 | 2 | 491 | 2.10 | 2/ 2 | 2.56 | 1/ 2 | 1.11 |
| Rev response element high affinity site | 9 | 2 | 2 | 4 | 3.95 | 1/ 2 | 4.42 | 0/ 2 | 1.96 |
| J4/5 from P4-P6 domain, Tetrahymena ribozyme |
9 | 2 | 1 | 335 | 1.76 | 1/ 2 | 2.12 | 1/ 2 | 1.09 |
| Tetraloop/helix interaction, L1 ligase crystal |
10 | 3 | 1 | 500 | 1.10 | 1/ 3 | 1.21 | 2/ 3 | 0.69 |
| Hook-turn motif | 11 | 3 | 5 | 121 | 2.56 | 3/ 3 | 2.06 | 3/ 3 | 1.37 |
| Helix with A-C base pairs | 12 | 2 | 2 | 242 | 2.45 | 1/ 4 | 1.81 | 2/ 4 | 1.53 |
| Curved helix with G-A and A-A base pairs |
12 | 2 | 1 | 205 | 1.74 | 2/ 4 | 1.06 | 4/ 4 | 0.96 |
| Fragment with G-G and G-A base pairs, SRP helix VI |
12 | 2 | 3 | 98 | 3.27 | 0/ 5 | 4.25 | 0/ 5 | 0.86 |
| Signal recognition particle Domain IV | 12 | 2 | 4 | 321 | 1.54 | 2/ 5 | 1.22 | 4/ 5 | 0.93 |
| Stem C internal loop, L1 ligase | 12 | 2 | 1 | 489 | 2.24 | 2/ 3 | 2.42 | 2/ 3 | 1.88 |
| Four-way junction, HCV IRES | 13 | 4 | 3 | 30 | 10.09 | 1/ 4 | 10.63 | 1/ 4 | 2.99 |
| Bulged G motif, sarcin/ricin loop | 13 | 2 | 1 | 81 | 1.46 | 4/ 4 | 1.66 | 3/ 4 | 0.86 |
| Kink-turn motif from SAM-I riboswitch | 13 | 2 | 1 | 7 | 1.43 | 3/ 3 | 1.36 | 3/ 3 | 1.22 |
| Three-way junction, purine riboswitch | 13 | 3 | 3 | 24 | 6.15 | 0/ 3 | 6.10 | 0/ 3 | 3.16 |
| J4a-4b region, metal-sensing riboswitch |
14 | 2 | 3 | 4 | 3.71 | 0/ 2 | 3.52 | 0/ 2 | 1.27 |
| Kink-turn motif | 15 | 2 | 2 | 25 | 8.85 | 1/ 3 | 9.43 | 2/ 3 | 3.05 |
| Tetraloop/receptor, P4-P6 domain, Tetr. Ribozyme |
15 | 3 | 4 | 13 | 3.31 | 2/ 5 | 2.89 | 2/ 5 | 2.21 |
| Tertiary interaction, hammerhead ribozyme |
16 | 3 | 2 | 4 | 7.82 | 0/ 3 | 8.50 | 1/ 3 | 4.37 |
| Active site, hammerhead ribozyme | 17 | 3 | 4 | 5 | 8.64 | 1/ 3 | 9.28 | 1/ 3 | 4.41 |
| J5-5a hinge, P4-P6 domain, Tetr. Ribozyme |
17 | 2 | 3 | 12 | 9.99 | 0/ 4 | 10.12 | 0/ 4 | 4.23 |
| Loop E motif, 5S RNA | 18 | 2 | 2 | 40 | 1.64 | 3/ 6 | 2.16 | 6/ 6 | 1.43 |
| L2-L3 tertiary interaction, purine riboswitch |
18 | 2 | 2 | 10 | 8.19 | 0/ 7 | 8.08 | 0/ 7 | 5.04 |
| Pseudoknot, domain III, CPV IRES | 18 | 2 | 4 | 11 | 3.55 | 0/ 0 | 3.90 | 0/ 0 | 2.29 |
| Pre-catalytic conformation, hammerhead ribozyme |
19 | 3 | 5 | 2 | 8.44 | 1/ 4 | 7.66 | 0/ 4 | 4.80 |
| P1-L3, SAM-II riboswitch | 23 | 2 | 5 | 5 | 7.40 | 0/ 1 | 7.47 | 0/ 1 | 3.99 |
Heavy-atom RMSD to crystal structure.
Number of non-Watson-Crick base pairs in crystal structure recovered in the model. Assignment of base pairing followed an automated method based on the RNAVIEW algorithm; counts of correct base pairings are lowered due to ambiguities in assigning bifurcated base pairs, pairs connected by single hydrogen bonds, or pairs that are not completely coplanar.