Table 9

Pfold predictions on both the high and medium similarity data-sets underflow on E. coli LSU rRNA. RNAalifold and Carnac, however, produce reasonable results.

E. coli LSU rRNA: Comparative Methods


Algorithm

number of bps in reference

number of bps in prediction

True Positives (% sensitivity)

False Positives (% selectivity)

Correlation (%)

Plan A: ClustalW Alignment

RNAalifold (H)

794

879

627 (79.0)

195 (76.3)

0.776 (77.6)

RNAalifold (H) + RNAfold-C

794

871

629 (79.2)

185 (77.3)

0.782 (78.2)

RNAalifold (M)

819

721

614 (75.0)

53 (92.1)

0.831 (83.5)

RNAalifold (M) + RNAfold-C

819

790

691 (84.4)

78 (89.9)

0.871 (87.1)

Pfold (H)

794

0

0 (0.0)

0 (0.0)

0.000 (0.0)

Pfold (M)

819

0

0 (0.0)

0 (0.0)

0.000 (0.0)

ILM (H)

794

1048

389 (49.0)

602 (39.3)

0.438 (44.1)

ILM (M)

819

1161

560 (68.4)

405 (58.0)

0.630 (63.2)

ILM-pknot (H)

869

1048

272 (31.3)

759 (26.4)

0.287 (28.8)

ILM-pknot (M)

869

1161

377 (43.4)

629 (37.5)

0.403 (40.4)

Plan B: Unaligned sequences

Carnac (H)

816

422

390 (47.8)

7 (98.2)

0.685 (73.0)

Carnac (H) + RNAfold-C

816

873

674 (82.6)

156 (81.2)

0.819 (81.9)

Carnac (M)

821

508

463 (56.4)

14 (97.1)

0.740 (76.7)

Carnac (M) + RNAfold-C

821

865

682 (83.1)

147 (82.3)

0.827 (82.7)


Gardner and Giegerich BMC Bioinformatics 2004 5:140   doi:10.1186/1471-2105-5-140

Open Data