Table 8.
RNA Contact Distance | ||||||||||
Frequency1 | 2–100 nt | 101–500 nt | 501+ nt | |||||||
750 | 21,049 | 98% | 18% | 417 | 2% | 2% | 0 | 0% | 0% | 21,466 |
600–749 | 42,362 | 94% | 37% | 2,805 | 6% | 14% | 33 | 0% | 2% | 45,200 |
151–599 | 20,775 | 81% | 18% | 4,594 | 18% | 23% | 266 | 1% | 18% | 25,635 |
1–150 | 31,285 | 70% | 27% | 12,253 | 27% | 61% | 1,161 | 3% | 80% | 44,699 |
Correct | 115,471 | 84% | 20,069 | 15% | 1,460 | 1% | 137,000 | |||
Never | 29,587 | 54% | 22,935 | 42% | 2,472 | 4% | 54,994 | |||
Total | 145,058 | 76% | 43,004 | 22% | 3,932 | 2% | 191,994 |
†For all 496 16S rRNA sequences, a total of 750 structure models are predicted for each sequence (one optimal and 749 suboptimal structure predictions). Every base-pair (excluding any base-pairs involving IUPAC symbols other than A,G,C, or U) in the comparative structure model that appears in a set of 750 structure predictions for a particular sequence is categorized by 1) the number of structure predictions in which it appears and 2) the RNA contact distance. The four bold percentages for each of the three RNA contact distances each total 100%, and reveal the percentage of base-pairs predicted correctly for the four frequency ranges. For example, a total of 115,471 base-pairs with an RNA contact distance of 2–100 nt were predicted correctly. Of those base-pairs, 18% (21,049) were predicted in 750 structure predictions, 37% (42,362) were predicted in 600–749 structure predictions, 18% (20,775) were predicted in 151–599 structure predictions, and 27% (31,285) were predicted in 1–150 structure predictions. In contrast, the three italicized percentages for each of the four frequency ranges, and the "Correct", "Never", and "Total" categories total 100%. For example, 54,994 base-pairs were never predicted in 750 structure predictions. Of those base-pairs, 54% (29,587) have an RNA contact distance of 2–100 nt, 42% (22,935) have an RNA contact distance of 101–500 nt, and 4% (2,472) have an RNA contact distance of 501+ nt.
1 Frequency of prediction throughout a suboptimal population of up to 750 structure predictions.