Accuracy Of Secondary Structure Prediction Research Articles

Computational approaches for RNA energy parameter estimation

Methods for efficient and accurate prediction of RNA structure are increasingly valuable, given the current rapid advances in understanding the diverse functions of RNA molecules in the cell. To enhance the accuracy of secondary structure predictions, we developed and refined optimization techniques for the estimation of energy parameters. We build on two previous approaches to RNA free-energy parameter estimation: (1) the Constraint Generation (CG) method, which iteratively generates constraints that enforce known structures to have energies lower than other structures for the same molecule; and (2) the Boltzmann Likelihood (BL) method, which infers a set of RNA free-energy parameters that maximize the conditional likelihood of a set of reference RNA structures. Here, we extend these approaches in two main ways: We propose (1) a max-margin extension of CG, and (2) a novel linear Gaussian Bayesian network that models feature relationships, which effectively makes use of sparse data by sharing statistical strength between parameters. We obtain significant improvements in the accuracy of RNA minimum free-energy pseudoknot-free secondary structure prediction when measured on a comprehensive set of 2518 RNA molecules with reference structures. Our parameters can be used in conjunction with software that predicts RNA secondary structures, RNA hybridization, or ensembles of structures. Our data, software, results, and parameter sets in various formats are freely available at http://www.cs.ubc.ca/labs/beta/Projects/RNA-Params.

Improving the accuracy of predicting secondary structure for aligned RNA sequences

Considerable attention has been focused on predicting the secondary structure for aligned RNA sequences since it is useful not only for improving the limiting accuracy of conventional secondary structure prediction but also for finding non-coding RNAs in genomic sequences. Although there exist many algorithms of predicting secondary structure for aligned RNA sequences, further improvement of the accuracy is still awaited. In this article, toward improving the accuracy, a theoretical classification of state-of-the-art algorithms of predicting secondary structure for aligned RNA sequences is presented. The classification is based on the viewpoint of maximum expected accuracy (MEA), which has been successfully applied in various problems in bioinformatics. The classification reveals several disadvantages of the current algorithms but we propose an improvement of a previously introduced algorithm (CentroidAlifold). Finally, computational experiments strongly support the theoretical classification and indicate that the improved CentroidAlifold substantially outperforms other algorithms.

Improvement of RNA secondary structure prediction using RNase H cleavage and randomized oligonucleotides

RNA secondary structure prediction using free energy minimization is one method to gain an approximation of structure. Constraints generated by enzymatic mapping or chemical modification can improve the accuracy of secondary structure prediction. We report a facile method that identifies single-stranded regions in RNA using short, randomized DNA oligonucleotides and RNase H cleavage. These regions are then used as constraints in secondary structure prediction. This method was used to improve the secondary structure prediction of Escherichia coli 5S rRNA. The lowest free energy structure without constraints has only 27% of the base pairs present in the phylogenetic structure. The addition of constraints from RNase H cleavage improves the prediction to 100% of base pairs. The same method was used to generate secondary structure constraints for yeast tRNAPhe, which is accurately predicted in the absence of constraints (95%). Although RNase H mapping does not improve secondary structure prediction, it does eliminate all other suboptimal structures predicted within 10% of the lowest free energy structure. The method is advantageous over other single-stranded nucleases since RNase H is functional in physiological conditions. Moreover, it can be used for any RNA to identify accessible binding sites for oligonucleotides or small molecules.

Correlation of RNA Secondary Structure Statistics with Thermodynamic Stability and Applications to Folding

The accurate prediction of the secondary and tertiary structure of an RNA with different folding algorithms is dependent on several factors, including the energy functions. However, an RNA higher-order structure cannot be predicted accurately from its sequence based on a limited set of energy parameters. The inter- and intramolecular forces between this RNA and other small molecules and macromolecules, in addition to other factors in the cell such as pH, ionic strength, and temperature, influence the complex dynamics associated with transition of a single stranded RNA to its secondary and tertiary structure. Since all of the factors that affect the formation of an RNAs 3D structure cannot be determined experimentally, statistically derived potential energy has been used in the prediction of protein structure. In the current work, we evaluate the statistical free energy of various secondary structure motifs, including base-pair stacks, hairpin loops, and internal loops, using their statistical frequency obtained from the comparative analysis of more than 50,000 RNA sequences stored in the RNA Comparative Analysis Database (rCAD) at the Comparative RNA Web (CRW) Site. Statistical energy was computed from the structural statistics for several datasets. While the statistical energy for a base-pair stack correlates with experimentally derived free energy values, suggesting a Boltzmann-like distribution, variation is observed between different molecules and their location on the phylogenetic tree of life. Our statistical energy values calculated for several structural elements were utilized in the Mfold RNA-folding algorithm. The combined statistical energy values for base-pair stacks, hairpins and internal loop flanks result in a significant improvement in the accuracy of secondary structure prediction; the hairpin flanks contribute the most.

Beyond the Twilight Zone: Automated prediction of structural properties of proteins by recursive neural networks and remote homology information

AbstractThe prediction of 1D structural properties of proteins is an important step toward the prediction of protein structure and function, not only in the ab initio case but also when homology information to known structures is available. Despite this the vast majority of 1D predictors do not incorporate homology information into the prediction process. We develop a novel structural alignment method, SAMD, which we use to build alignments of putative remote homologues that we compress into templates of structural frequency profiles. We use these templates as additional input to ensembles of recursive neural networks, which we specialise for the prediction of query sequences that show only remote homology to any Protein Data Bank structure. We predict four 1D structural properties – secondary structure, relative solvent accessibility, backbone structural motifs, and contact density. Secondary structure prediction accuracy, tested by five‐fold cross‐validation on a large set of proteins allowing less than 25% sequence identity between training and test set and query sequences and templates, exceeds 82%, outperforming its ab initio counterpart, other state‐of‐the‐art secondary structure predictors (Jpred 3 and PSIPRED) and two other systems based on PSI‐BLAST and COMPASS templates. We show that structural information from homologues improves prediction accuracy well beyond the Twilight Zone of sequence similarity, even below 5% sequence identity, for all four structural properties. Significant improvement over the extraction of structural information directly from PDB templates suggests that the combination of sequence and template information is more informative than templates alone. Proteins 2009. © 2009 Wiley‐Liss, Inc.

Prediction of RNA secondary structure using generalized centroid estimators

Recent studies have shown that the methods for predicting secondary structures of RNAs on the basis of posterior decoding of the base-pairing probabilities has an advantage with respect to prediction accuracy over the conventionally utilized minimum free energy methods. However, there is room for improvement in the objective functions presented in previous studies, which are maximized in the posterior decoding with respect to the accuracy measures for secondary structures. We propose novel estimators which improve the accuracy of secondary structure prediction of RNAs. The proposed estimators maximize an objective function which is the weighted sum of the expected number of the true positives and that of the true negatives of the base pairs. The proposed estimators are also improved versions of the ones used in previous works, namely CONTRAfold for secondary structure prediction from a single RNA sequence and McCaskill-MEA for common secondary structure prediction from multiple alignments of RNA sequences. We clarify the relations between the proposed estimators and the estimators presented in previous works, and theoretically show that the previous estimators include additional unnecessary terms in the evaluation measures with respect to the accuracy. Furthermore, computational experiments confirm the theoretical analysis by indicating improvement in the empirical accuracy. The proposed estimators represent extensions of the centroid estimators proposed in Ding et al. and Carvalho and Lawrence, and are applicable to a wide variety of problems in bioinformatics. Supporting information and the CentroidFold software are available online at: http://www.ncrna.org/software/centroidfold/.

Combining hydrophobicity with PSSM for protein secondary structure prediction using BP neural network

Objective Since predicting protein secondary structure is the basis of predicting protein spacial structure, it is important to improve the prediction accuracy of secondary structure. Methods A two-stage BP neural network was constructed based on the method of combining hydrophobicity of amino acid residues with PSSM which contains evolution information. CB513 dataset was employed in our study. After excluding the protein chains containing X,B and those with sequence length shorter than 30 amino acids, 492 protein chains in the dataset were used. The results of protein secondary structure prediction of our study were compared with those from the networks using only PSSN as their inputs. Accuracy of the network was tested by 4-fold cross-validation. Results In our study, α-helix was predicted with an averaged accuracy of nearly 79%, sensitivity of 79% and specificity of 91%. Total prediction accuracy of secondary structure reached 75.96%, which was higher than that of only using PSSM as input. Conclusion The new method developed can better predict protein secondary structure, especially α-helix with a higher accuracy. Key words: Protein secondary structure; Prediction; BP neural network; Hydrophobicity; Positionspecific scoring matrix

Impact of residue accessible surface area on the prediction of protein secondary structures

BackgroundThe problem of accurate prediction of protein secondary structure continues to be one of the challenging problems in Bioinformatics. It has been previously suggested that amino acid relative solvent accessibility (RSA) might be an effective factor for increasing the accuracy of protein secondary structure prediction. Previous studies have either used a single constant threshold to classify residues into discrete classes (buries vs. exposed), or used the real-value predicted RSAs in their prediction method.ResultsWe studied the effect of applying different RSA threshold types (namely, fixed thresholds vs. residue-dependent thresholds) on a variety of secondary structure prediction methods. With the consideration of DSSP-assigned RSA values we realized that improvement in the accuracy of prediction strictly depends on the selected threshold(s). Furthermore, we showed that choosing a single threshold for all amino acids is not the best possible parameter. We therefore used residue-dependent thresholds and most of residues showed improvement in prediction. Next, we tried to consider predicted RSA values, since in the real-world problem, protein sequence is the only available information. We first predicted the RSA classes by RVP-net program and then used these data in our method. Using this approach, improvement in prediction was also obtained.ConclusionThe success of applying the RSA information on different secondary structure prediction methods suggest that prediction accuracy can be improved independent of prediction approaches. Thus, solvent accessibility can be considered as a rich source of information to help the improvement of these methods.

Thermodynamic Characterization of the Complete Set of Sequence Symmetric Tandem Mismatches in RNA and an Improved Model for Predicting the Free Energy Contribution of Sequence Asymmetric Tandem Mismatches

Because of the availability of an abundance of RNA sequence information, the ability to rapidly and accurately predict the secondary structure of RNA from sequence is becoming increasingly important. A common method for predicting RNA secondary structure from sequence is free energy minimization. Therefore, accurate free energy contributions for every RNA secondary structure motif are necessary for accurate secondary structure predictions. Tandem mismatches are prevalent in naturally occurring sequences and are biologically important. A common method for predicting the stability of a sequence asymmetric tandem mismatch relies on the stabilities of the two corresponding sequence symmetric tandem mismatches [Mathews, D. H., Sabina, J., Zuker, M., and Turner, D. H. (1999) J. Mol. Biol. 288, 911-940]. To improve the prediction of sequence asymmetric tandem mismatches, the experimental thermodynamic parameters for the 22 previously unmeasured sequence symmetric tandem mismatches are reported. These new data, however, do not improve prediction of the free energy contributions of sequence asymmetric tandem mismatches. Therefore, a new model, independent of sequence symmetric tandem mismatch free energies, is proposed. This model consists of two penalties to account for destabilizing tandem mismatches, two bonuses to account for stabilizing tandem mismatches, and two penalties to account for A-U and G-U adjacent base pairs. This model improves the prediction of asymmetric tandem mismatch free energy contributions and is likely to improve the prediction of RNA secondary structure from sequence.

TMBpro: secondary structure, β-contact and tertiary structure prediction of transmembrane β-barrel proteins

Transmembrane beta-barrel (TMB) proteins are embedded in the outer membranes of mitochondria, Gram-negative bacteria and chloroplasts. These proteins perform critical functions, including active ion-transport and passive nutrient intake. Therefore, there is a need for accurate prediction of secondary and tertiary structure of TMB proteins. Traditional homology modeling methods, however, fail on most TMB proteins since very few non-homologous TMB structures have been determined. Yet, because TMB structures conform to specific construction rules that restrict the conformational space drastically, it should be possible for methods that do not depend on target-template homology to be applied successfully. We develop a suite (TMBpro) of specialized predictors for predicting secondary structure (TMBpro-SS), beta-contacts (TMBpro-CON) and tertiary structure (TMBpro-3D) of transmembrane beta-barrel proteins. We compare our results to the recent state-of-the-art predictors transFold and PRED-TMBB using their respective benchmark datasets, and leave-one-out cross-validation. Using the transFold dataset TMBpro predicts secondary structure with per-residue accuracy (Q(2)) of 77.8%, a correlation coefficient of 0.54, and TMBpro predicts beta-contacts with precision of 0.65 and recall of 0.67. Using the PRED-TMBB dataset, TMBpro predicts secondary structure with Q(2) of 88.3% and a correlation coefficient of 0.75. All of these performance results exceed previously published results by 4% or more. Working with the PRED-TMBB dataset, TMBpro predicts the tertiary structure of transmembrane segments with RMSD <6.0 A for 9 of 14 proteins. For 6 of 14 predictions, the RMSD is <5.0 A, with a GDT_TS score greater than 60.0. http://www.igb.uci.edu/servers/psss.html.

Base pairing in RNA structures: A computational analysis of structural aspects and interaction energies

The base pairing patterns in RNA structures are more versatile and completely different as compared to DNA. We present here results of ab-initio studies of structures and interaction energies of eight selected RNA base pairs reported in literature. Interaction energies, including BSSE correction, of hydrogen added crystal geometries of base pairs have been calculated at the HF/6-31G** level. The structures and interaction energies of the base pairs in the crystal geometry are compared with those obtained after optimization of the base pairs. We find that the base pairs become more planar on full optimization. No change in the hydrogen bonding pattern is seen. It is expected that the inclusion of appropriate considerations of many of these aspects of RNA base pairing would significantly improve the accuracy of RNA secondary structure prediction.

Use of secondary structural information and Cα-Cα distance restraints to model protein structures with MODELLER

Protein secondary structure predictions and amino acid long range contact map predictions from primary sequence of proteins have been explored to aid in modelling protein tertiary structures. In order to evaluate the usefulness of secondary structure and 3D-residue contact prediction methods to model protein structures we have used the known Q3 (alpha-helix,beta-strands and irregular turns/loops) secondary structure information, along with residue-residue contact information as restraints for MODELLER. We present here results of our modelling studies on 30 best resolved single domain protein structures of varied lengths. The results shows that it is very difficult to obtain useful models even with 100% accurate secondary structure predictions and accurate residue contact predictions for up to 30% of residues in a sequence. The best models that we obtained for proteins of lengths 37, 70, 118, 136 and 193 amino acid residues are of RMSDs 4.17, 5.27, 9.12, 7.89 and 9.69,respectively. The results show that one can obtain better models for the proteins which have high percent of alpha-helix content. This analysis further shows that MODELLER restrain optimization program can be useful only if we have truly homologous structure(s) as a template where it derives numerous restraints, almost identical to the templates used. This analysis also clearly indicates that even if we satisfy several true residue-residue contact distances, up to 30%of their sequence length with fully known secondary structural information, we end up predicting model structures much distant from their corresponding native structures.

Consensus Data Mining (CDM) Protein Secondary Structure Prediction Server: Combining GOR V and Fragment Database Mining (FDM)

One of the challenges in protein secondary structure prediction is to overcome the cross-validated 80% prediction accuracy barrier. Here, we propose a novel approach to surpass this barrier. Instead of using a single algorithm that relies on a limited data set for training, we combine two complementary methods having different strengths: Fragment Database Mining (FDM) and GOR V. FDM harnesses the availability of the known protein structures in the Protein Data Bank and provides highly accurate secondary structure predictions when sequentially similar structural fragments are identified. In contrast, the GOR V algorithm is based on information theory, Bayesian statistics, and PSI-BLAST multiple sequence alignments to predict the secondary structure of residues inside a sliding window along a protein chain. A combination of these two different methods benefits from the large number of structures in the PDB and significantly improves the secondary structure prediction accuracy, resulting in Q3 ranging from 67.5 to 93.2%, depending on the availability of highly similar fragments in the Protein Data Bank.

Efficient parameter estimation for RNA secondary structure prediction

Accurate prediction of RNA secondary structure from the base sequence is an unsolved computational challenge. The accuracy of predictions made by free energy minimization is limited by the quality of the energy parameters in the underlying free energy model. The most widely used model, the Turner99 model, has hundreds of parameters, and so a robust parameter estimation scheme should efficiently handle large data sets with thousands of structures. Moreover, the estimation scheme should also be trained using available experimental free energy data in addition to structural data. In this work, we present constraint generation (CG), the first computational approach to RNA free energy parameter estimation that can be efficiently trained on large sets of structural as well as thermodynamic data. Our CG approach employs a novel iterative scheme, whereby the energy values are first computed as the solution to a constrained optimization problem. Then the newly computed energy parameters are used to update the constraints on the optimization function, so as to better optimize the energy parameters in the next iteration. Using our method on biologically sound data, we obtain revised parameters for the Turner99 energy model. We show that by using our new parameters, we obtain significant improvements in prediction accuracy over current state of-the-art methods. Our CG implementation is available at http://www.rnasoft.ca/CG/.

Boltzmann probability of RNA structural neighbors and riboswitch detection

We describe algorithms implemented in a new software package, RNAbor, to investigate structures in a neighborhood of an input secondary structure S of an RNA sequence s. The input structure could be the minimum free energy structure, the secondary structure obtained by analysis of the X-ray structure or by comparative sequence analysis, or an arbitrary intermediate structure. A secondary structure T of s is called a delta-neighbor of S if T and S differ by exactly delta base pairs. RNAbor computes the number (N(delta)), the Boltzmann partition function (Z(delta)) and the minimum free energy (MFE(delta)) and corresponding structure over the collection of all delta-neighbors of S. This computation is done simultaneously for all delta < or = m, in run time O (mn3) and memory O(mn2), where n is the sequence length. We apply RNAbor for the detection of possible RNA conformational switches, and compare RNAbor with the switch detection method paRNAss. We also provide examples of how RNAbor can at times improve the accuracy of secondary structure prediction. http://bioinformatics.bc.edu/clotelab/RNAbor/. Supplementary data are available at Bioinformatics online.

Murlet: a practical multiple alignment tool for structural RNA sequences

Structural RNA genes exhibit unique evolutionary patterns that are designed to conserve their secondary structures; these patterns should be taken into account while constructing accurate multiple alignments of RNA genes. The Sankoff algorithm is a natural alignment algorithm that includes the effect of base-pair covariation in the alignment model. However, the extremely high computational cost of the Sankoff algorithm precludes its application to most RNA sequences. We propose an efficient algorithm for the multiple alignment of structural RNA sequences. Our algorithm is a variant of the Sankoff algorithm, and it uses an efficient scoring system that reduces the time and space requirements considerably without compromising on the alignment quality. First, our algorithm computes the match probability matrix that measures the alignability of each position pair between sequences as well as the base pairing probability matrix for each sequence. These probabilities are then combined to score the alignment using the Sankoff algorithm. By itself, our algorithm does not predict the consensus secondary structure of the alignment but uses external programs for the prediction. We demonstrate that both the alignment quality and the accuracy of the consensus secondary structure prediction from our alignment are the highest among the other programs examined. We also demonstrate that our algorithm can align relatively long RNA sequences such as the eukaryotic-type signal recognition particle RNA that is approximately 300 nt in length; multiple alignment of such sequences has not been possible by using other Sankoff-based algorithms. The algorithm is implemented in the software named 'Murlet'. The C++ source code of the Murlet software and the test dataset used in this study are available at http://www.ncrna.org/papers/Murlet/. Supplementary data are available at Bioinformatics online.

Achieving 80% ten‐fold cross‐validated accuracy for secondary structure prediction by large‐scale training

An integrated system of neural networks, called SPINE, is established and optimized for predicting structural properties of proteins. SPINE is applied to three-state secondary-structure and residue-solvent-accessibility (RSA) prediction in this paper. The integrated neural networks are carefully trained with a large dataset of 2640 chains, sequence profiles generated from multiple sequence alignment, representative amino acid properties, a slow learning rate, overfitting protection, and an optimized sliding-widow size. More than 200,000 weights in SPINE are optimized by maximizing the accuracy measured by Q(3) (the percentage of correctly classified residues). SPINE yields a 10-fold cross-validated accuracy of 79.5% (80.0% for chains of length between 50 and 300) in secondary-structure prediction after one-month (CPU time) training on 22 processors. An accuracy of 87.5% is achieved for exposed residues (RSA >95%). The latter approaches the theoretical upper limit of 88-90% accuracy in assigning secondary structures. An accuracy of 73% for three-state solvent-accessibility prediction (25%/75% cutoff) and 79.3% for two-state prediction (25% cutoff) is also obtained.

PROMALS: towards accurate multiple sequence alignments of distantly related proteins

Accurate multiple sequence alignments are essential in protein structure modeling, functional prediction and efficient planning of experiments. Although the alignment problem has attracted considerable attention, preparation of high-quality alignments for distantly related sequences remains a difficult task. We developed PROMALS, a multiple alignment method that shows promising results for protein homologs with sequence identity below 10%, aligning close to half of the amino acid residues correctly on average. This is about three times more accurate than traditional pairwise sequence alignment methods. PROMALS algorithm derives its strength from several sources: (i) sequence database searches to retrieve additional homologs; (ii) accurate secondary structure prediction; (iii) a hidden Markov model that uses a novel combined scoring of amino acids and secondary structures; (iv) probabilistic consistency-based scoring applied to progressive alignment of profiles. Compared to the best alignment methods that do not use secondary structure prediction and database searches (e.g. MUMMALS, ProbCons and MAFFT), PROMALS is up to 30% more accurate, with improvement being most prominent for highly divergent homologs. Compared to SPEM and HHalign, which also employ database searches and secondary structure prediction, PROMALS shows an accuracy improvement of several percent. The PROMALS web server is available at: http://prodata.swmed.edu/promals/. Supplementary data are available at Bioinformatics online.

A new expertness index for assessment of secondary structure prediction engines

Improvement of prediction accuracy of the protein secondary structure is essential for further developments of the whole field of protein research. In this paper, the expertness of protein secondary structure prediction engines has been studied in three levels and a new criterion has been introduced in the third level. This criterion could be considered as an extension of the previous ones based on amino acid index. Using this new criterion, the expertness of some high score secondary structure prediction engines has been reanalyzed and some hidden facts have been discovered. The results of this new assessment demonstrated that a noticeable harmony has been existed among each amino acid prediction behavior in all engines. This harmony has also been seen between single global propensity and prediction accuracy of amino acid types in each secondary structure class. Moreover, it is shown that Proline and Glycine amino acids have been predicted with less accuracy in alpha helices and beta strands. In addition, regardless of different approaches used in prediction engines, beta strands have been predicted with less accuracy.

Robust prediction of consensus secondary structures using averaged base pairing probability matrices

Recent transcriptomic studies have revealed the existence of a considerable number of non-protein-coding RNA transcripts in higher eukaryotic cells. To investigate the functional roles of these transcripts, it is of great interest to find conserved secondary structures from multiple alignments on a genomic scale. Since multiple alignments are often created using alignment programs that neglect the special conservation patterns of RNA secondary structures for computational efficiency, alignment failures can cause potential risks of overlooking conserved stem structures. We investigated the dependence of the accuracy of secondary structure prediction on the quality of alignments. We compared three algorithms that maximize the expected accuracy of secondary structures as well as other frequently used algorithms. We found that one of our algorithms, called McCaskill-MEA, was more robust against alignment failures than others. The McCaskill-MEA method first computes the base pairing probability matrices for all the sequences in the alignment and then obtains the base pairing probability matrix of the alignment by averaging over these matrices. The consensus secondary structure is predicted from this matrix such that the expected accuracy of the prediction is maximized. We show that the McCaskill-MEA method performs better than other methods, particularly when the alignment quality is low and when the alignment consists of many sequences. Our model has a parameter that controls the sensitivity and specificity of predictions. We discussed the uses of that parameter for multi-step screening procedures to search for conserved secondary structures and for assigning confidence values to the predicted base pairs. The C++ source code that implements the McCaskill-MEA algorithm and the test dataset used in this paper are available at http://www.ncrna.org/papers/McCaskillMEA/. Supplementary data are available at Bioinformatics online.