Improved prediction of RNA secondary structure by integrating the free energy model with restraints derived from experimental probing data.
Bottom Line: We first demonstrated that RME substantially improved secondary structure prediction with perfect restraints (base pair information of known structures).For high-throughput data (e.g. PARS and DMS-seq) with lower probing efficiency, the secondary structure prediction performances of the tested tools were comparable, with performance improvements for only a portion of the tested RNAs.However, when the effects of tertiary structure and protein interactions were removed, RME showed the highest prediction accuracy in the DMS-accessible regions by incorporating in vivo DMS-seq data.
Affiliation: MOE Key Laboratory of Bioinformatics, Center for Synthetic and Systems Biology, Center for Plant Biology and Tsinghua-Peking Joint Center for Life Sciences, School of Life Sciences, Tsinghua University, Beijing 100084, China.Show MeSH
Related in: MedlinePlus
Mentions: The average performance scores for the RNA secondary structure predictions from the five-fold cross-validation are shown in Figure 2 (details in Supplementary Table S7). RME with perfect restraints significantly improved the accuracy of RNA secondary structure prediction in comparison with RME-control without restraints added. We compared the mean and standard deviation of the average MCC from five-fold cross validation. The average MCC was increased from (62.9 ± 0.6)% (sample size = 5) for RME-control to (93.7 ± 0.2)% (sample size = 5) for RME (P < 0.05, one-tailed Wilcoxon signed-rank test). RNAstructure-Fold also performed well with perfect restraints: the average MCC was increased from (62.0 ± 0.5)% (sample size = 5) to (93.4 ± 0.2)% (sample size = 5). Although the addition of perfect restraints to SeqFold (average MCC, (68.4 ± 0.7)%, sample size = 5) produced great improvement in comparison with SeqFold-control (average MCC, (61.6 ± 0.7)%, sample size = 5), SeqFold did not perform as well as RME or RNAstucture-Fold, because SeqFold cannot guarantee the correct structure is sampled, even when the restraint data is perfect.
Affiliation: MOE Key Laboratory of Bioinformatics, Center for Synthetic and Systems Biology, Center for Plant Biology and Tsinghua-Peking Joint Center for Life Sciences, School of Life Sciences, Tsinghua University, Beijing 100084, China.