Detection of non-coding RNAs on the basis of predicted secondary structure formation free energy change

BMC Bioinformatics

Table 1 Percent of known base pairs in predicted suboptimal structures for single sequences.

	Maximum percent change in free energy from lowest free energy structure
RNA Type¹	1%	5%	10%	20%	30%	50%
SSU (16S) rRNA	74.5 ± 21.9 (80.5 ± 16.0)²	88.1 ± 14.9 (96.8 ± 2.7)²	97.1 ± 13.6 (97.2 ± 1.4)²	99.2 ± 8.2 (97.2 ± 1.4)²	99.3 ± 7.2 (97.2 ± 1.4)²	99.3 ± 3.4 (97.2 ± 1.4)²
LSU (23S) rRNA	84.4 ± 8.9 (91.9 ± 13.4)²	96.8 ± 3.8 (97.9 ± 1.2)²	98.1 ± 1.2 (98.0 ± 0.7)²	98.1 ± 1.2 (98.0 ± 0.7)²	98.1 ± 1.2 (98.0 ± 0.7)²	98.1 ± 1.2 (98.0 ± 0.7)²
5S rRNA	74.5 ± 25.6	88.1 ± 20.1	97.1 ± 7.4	99.2 ± 1.6	99.3 ± 1.4	99.3 ± 1.4
Group I Intron	79.0 ± 12.6	93.5 ± 7.6	98.3 ± 1.4	98.4 ± 1.4	98.4 ± 1.4	98.4 ± 1.4
Group I Intron – 2	74.4 ± 13.6	92.8 ± 8.0	97.1 ± 1.7	97.1 ± 1.7	97.1 ± 1.7	97.1 ± 1.7
Group II Intron	91.9 ± 5.7	100.0 ± 0.0	100.0 ± 0.0	100.0 ± 0.0	100.0 ± 0.0	100.0 ± 0.0
RNase P	79.8 ± 11.0	95.5 ± 2.6	98.4 ± 1.2	98.4 ± 1.2	98.4 ± 1.2	98.4 ± 1.2
RNase P – 2	75.0 ± 7.9	95.6 ± 4.9	98.3 ± 1.5	98.3 ± 1.5	98.3 ± 1.5	98.3 ± 1.5
SRP RNA	73.1 ± 25.2	90.7 ± 14.3	95.5 ± 8.2	97.1 ± 2.7	97.2 ± 2.6	97.2 ± 2.6
tRNA	87.0 ± 18.2	94.5 ± 13.3	97.9 ± 7.8	99.3 ± 4.6	99.6 ± 3.2	99.8 ± 1.0
Average ³	80.5 ± 6.7	93.4 ± 4.2	97.8 ± 1.3	98.7 ± 0.9	98.8 ± 0.9	98.8 ± 1.0

¹The database of structures was assembled for previous studies of secondary structure prediction [48] and is derived from a diverse set of databases [76, 77, 81-84].
²The large and small subunit rRNA sequences are divided into domains of less than 700 nucleotides for structure prediction. In parenthesis are the accuracies when the whole sequence is folded at once.
³The average is calculated excluding the second database of Group II introns and RNase P sequences as was done in [48].
The percent of known base pairs contained in at least one predicted suboptimal structure within a specified percent difference in free energy from the minimum free energy. For example, 99.2% of known base pairs in 5S rRNA secondary structures are found on average within the predicted structures with less than 20% difference in free energy from the lowest free energy structure. These numbers were calculated from an energy dot plot using the thermodynamic parameters of [49]. On average, 98.7% of known base pairs occur in at least one suboptimal secondary structure within 20% or less difference in free energy from the minimum free energy structure. This accuracy remains similar as the percent energy difference is increased to 30% or 50%.

ISSN: 1471-2105