Overall | Archaea | Bacteria | Eucarya | |||
---|---|---|---|---|---|---|

(c)^{1} |
(m) | (n) | ||||

Comparative | 191,994 | 10,211 | 83,385 | 13,406 | 29,979 | 55,013 |

Opt Correct^{2} |
81,934 | 6,376 | 41,032 | 6,105 | 9,459 | 18,962 |

Subopt Correct^{3} |
137,000 | 8,570 | 65,177 | 10,032 | 21,201 | 32,020 |

Opt Incorrect^{2} |
142,023 | 4,758 | 49,563 | 8,603 | 27,617 | 51,482 |

Subopt Incorrect^{3} |
2,372,305 | 101,253 | 947,197 | 161,397 | 472,614 | 689,844 |

Opt Accuracy^{2,4} |
41% | 62% | 49% | 46% | 30% | 34% |

Subopt Accuracy^{3,4} |
71% | 84% | 78% | 75% | 71% | 59% |

Avg Improvement^{5} |
30% | 21% | 29% | 30% | 41% | 24% |

Best Prediction^{6} |
92% | 91% | 89% | 92% | 92% | 90% |

Max Improvement^{7} |
68% | 35% | 54% | 53% | 68% | 48% |

Min Improvement^{8} |
10% | 10% | 12% | 12% | 14% | 11% |

^{} All 496 16S rRNA sequences are considered. Each sequence is folded for a population of one optimal and 749 suboptimal structure predictions. The determination of the accuracy for the structures predicted with Mfold is described in the **Methods** section, *RNA Secondary Structure Prediction* and *Prediction Accuracy Calculations*. Values are calculated by summing the number of unique base-pairs encountered for each sequence that satisfy each particular category (any base-pairs involving IUPAC symbols other than A,G,C, or U are excluded). For example, *Subopt Correct* is calculated by summing the number of unique, correctly predicted base-pairs encountered in the population of optimal plus suboptimal structure predictions for each of the 496 16S rRNA sequences. Prediction accuracy when including base-pairs predicted correctly in suboptimal structure predictions is also tabulated.

- (c), Chloroplast-encoded sequences; (m), Mitochondrial-encoded sequences; (n), Nuclear-encoded sequences.
- Considering only the optimal prediction.
- Considering the optimal prediction plus up to 749 suboptimal predictions.
- Averages calculated on per sequence basis. Please see
*Per Sequence Averages*in**Methods**. - Average improvement in MFold secondary structure prediction accuracy when pooling base-pairs from both the optimal prediction and suboptimal predictions.
- The highest MFold secondary structure prediction accuracy for an individual sequence when pooling base-pairs from both the optimal and suboptimal populations.
- The largest improvement in MFold secondary structure prediction accuracy for an individual sequence when pooling base-pairs from both the optimal and suboptimal populations.
- The smallest improvement in MFold secondary structure prediction accuracy for an individual sequence when pooling base-pairs from both the optimal and suboptimal populations.