CONTRAfold: RNA secondary structure prediction without physics-based models

MOTIVATION: For several decades, free energy minimization methods have been the dominant strategy for single sequence RNA secondary structure prediction. More recently, stochastic context-free grammars (SCFGs) have emerged as an alternative probabilistic methodology for modeling RNA structure. Unlike physics-based methods, which rely on thousands of experimentally-measured thermodynamic parameters, SCFGs use fully-automated statistical learning algorithms to derive model parameters. Despite this advantage, however, probabilistic methods have not replaced free energy minimization methods as the tool of choice for secondary structure prediction, as the accuracies of the best current SCFGs have yet to match those of the best physics-based models. RESULTS: In this paper, we present CONTRAfold, a novel secondary structure prediction method based on conditional log-linear models (CLLMs), a flexible class of probabilistic models which generalize upon SCFGs by using discriminative training and feature-rich scoring. In a series of cross-validation experiments, we show that grammar-based secondary structure prediction methods formulated as CLLMs consistently outperform their SCFG analogs. Furthermore, CONTRAfold, a CLLM incorporating most of the features found in typical thermodynamic models, achieves the highest single sequence prediction accuracies to date, outperforming currently available probabilistic and physics-based techniques. Our result thus closes the gap between probabilistic and thermodynamic models, demonstrating that statistical learning procedures provide an effective alternative to empirical measurement of thermodynamic parameters for RNA secondary structure prediction. AVAILABILITY: Source code for CONTRAfold is available at http://contra.stanford.edu/contrafold/.     

Evaluation of several lightweight stochastic context-free grammars for RNA secondary structure prediction

Background

RNA secondary structure prediction methods based on probabilistic modeling can be developed using stochastic context-free grammars (SCFGs). Such methods can readily combine different sources of information that can be expressed probabilistically, such as an evolutionary model of comparative RNA sequence analysis and a biophysical model of structure plausibility. However, the number of free parameters in an integrated model for consensus RNA structure prediction can become untenable if the underlying SCFG design is too complex. Thus a key question is, what small, simple SCFG designs perform best for RNA secondary structure prediction?

Results

Nine different small SCFGs were implemented to explore the tradeoffs between model complexity and prediction accuracy. Each model was tested for single sequence structure prediction accuracy on a benchmark set of RNA secondary structures.

Conclusions

Four SCFG designs had prediction accuracies near the performance of current energy minimization programs. One of these designs, introduced by Knudsen and Hein in their PFOLD algorithm, has only 21 free parameters and is significantly simpler than the others.

     

RNA secondary structure prediction using stochastic context-free grammars and evolutionary history.

MOTIVATION: Many computerized methods for RNA secondary structure prediction have been developed. Few of these methods, however, employ an evolutionary model, thus relevant information is often left out from the structure determination. This paper introduces a method which incorporates evolutionary history into RNA secondary structure prediction. The method reported here is based on stochastic context-free grammars (SCFGs) to give a prior probability distribution of structures. RESULTS: The phylogenetic tree relating the sequences can be found by maximum likelihood (ML) estimation from the model introduced here. The tree is shown to reveal information about the structure, due to mutation patterns. The inclusion of a prior distribution of RNA structures ensures good structure predictions even for a small number of related sequences. Prediction is carried out using maximum a posteriori estimation (MAP) estimation in a Bayesian approach. For small sequence sets, the method performs very well compared to current automated methods.     

Stochastic context-free grammars for tRNA modeling.

Stochastic context-free grammars (SCFGs) are applied to the problems of folding, aligning and modeling families of tRNA sequences. SCFGs capture the sequences' common primary and secondary structure and generalize the hidden Markov models (HMMs) used in related work on protein and DNA. Results show that after having been trained on as few as 20 tRNA sequences from only two tRNA subfamilies (mitochondrial and cytoplasmic), the model can discern general tRNA from similar-length RNA sequences of other kinds, can find secondary structure of new tRNA sequences, and can produce multiple alignments of large sets of tRNA sequences. Our results suggest potential improvements in the alignments of the D- and T-domains in some mitochondrial tRNAs that cannot be fit into the canonical secondary structure.     

Accelerated probabilistic inference of RNA structure evolution

Background  

Pairwise stochastic context-free grammars (Pair SCFGs) are powerful tools for evolutionary analysis of RNA, including simultaneous RNA sequence alignment and secondary structure prediction, but the associated algorithms are intensive in both CPU and memory usage. The same problem is faced by other RNA alignment-and-folding algorithms based on Sankoff's 1985 algorithm. It is therefore desirable to constrain such algorithms, by pre-processing the sequences and using this first pass to limit the range of structures and/or alignments that can be considered.     

Semantics and ambiguity of stochastic RNA family models

Stochastic models, such as hidden Markov models or stochastic context-free grammars (SCFGs) can fail to return the correct, maximum likelihood solution in the case of semantic ambiguity. This problem arises when the algorithm implementing the model inspects the same solution in different guises. It is a difficult problem in the sense that proving semantic nonambiguity has been shown to be algorithmically undecidable, while compensating for it (by coalescing scores of equivalent solutions) has been shown to be NP-hard. For stochastic context-free grammars modeling RNA secondary structure, it has been shown that the distortion of results can be quite severe. Much less is known about the case when stochastic context-free grammars model the matching of a query sequence to an implicit consensus structure for an RNA family. We find that three different, meaningful semantics can be associated with the matching of a query against the model--a structural, an alignment, and a trace semantics. Rfam models correctly implement the alignment semantics, and are ambiguous with respect to the other two semantics, which are more abstract. We show how provably correct models can be generated for the trace semantics. For approaches, where such a proof is not possible, we present an automated pipeline to check post factum for ambiguity of the generated models. We propose that both the structure and the trace semantics are worth-while concepts for further study, possibly better suited to capture remotely related family members.     

Evaluation of a sophisticated SCFG design for RNA secondary structure prediction

Predicting secondary structures of RNA molecules is one of the fundamental problems of and thus a challenging task in computational structural biology. Over the past decades, mainly two different approaches have been considered to compute predictions of RNA secondary structures from a single sequence: the first one relies on physics-based and the other on probabilistic RNA models. Particularly, the free energy minimization (MFE) approach is usually considered the most popular and successful method. Moreover, based on the paradigm-shifting work by McCaskill which proposes the computation of partition functions (PFs) and base pair probabilities based on thermodynamics, several extended partition function algorithms, statistical sampling methods and clustering techniques have been invented over the last years. However, the accuracy of the corresponding algorithms is limited by the quality of underlying physics-based models, which include a vast number of thermodynamic parameters and are still incomplete. The competing probabilistic approach is based on stochastic context-free grammars (SCFGs) or corresponding generalizations, like conditional log-linear models (CLLMs). These methods abstract from free energies and instead try to learn about the structural behavior of the molecules by learning (a manageable number of) probabilistic parameters from trusted RNA structure databases. In this work, we introduce and evaluate a sophisticated SCFG design that mirrors state-of-the-art physics-based RNA structure prediction procedures by distinguishing between all features of RNA that imply different energy rules. This SCFG actually serves as the foundation for a statistical sampling algorithm for RNA secondary structures of a single sequence that represents a probabilistic counterpart to the sampling extension of the PF approach. Furthermore, some new ways to derive meaningful structure predictions from generated sample sets are presented. They are used to compare the predictive accuracy of our model to that of other probabilistic and energy-based prediction methods. Particularly, comparisons to lightweight SCFGs and corresponding CLLMs for RNA structure prediction indicate that more complex SCFG designs might yield higher accuracy but eventually require more comprehensive and pure training sets. Investigations on both the accuracies of predicted foldings and the overall quality of generated sample sets (especially on an abstraction level, called abstract shapes of generated structures, that is relevant for biologists) yield the conclusion that the Boltzmann distribution of the PF sampling approach is more centered than the ensemble distribution induced by the sophisticated SCFG model, which implies a greater structural diversity within generated samples. In general, neither of the two distinct ensemble distributions is more adequate than the other and the corresponding results obtained by statistical sampling can be expected to bare fundamental differences, such that the method to be preferred for a particular input sequence strongly depends on the considered RNA type.     

A range of complex probabilistic models for RNA secondary structure prediction that includes the nearest-neighbor model and more

The standard approach for single-sequence RNA secondary structure prediction uses a nearest-neighbor thermodynamic model with several thousand experimentally determined energy parameters. An attractive alternative is to use statistical approaches with parameters estimated from growing databases of structural RNAs. Good results have been reported for discriminative statistical methods using complex nearest-neighbor models, including CONTRAfold, Simfold, and ContextFold. Little work has been reported on generative probabilistic models (stochastic context-free grammars [SCFGs]) of comparable complexity, although probabilistic models are generally easier to train and to use. To explore a range of probabilistic models of increasing complexity, and to directly compare probabilistic, thermodynamic, and discriminative approaches, we created TORNADO, a computational tool that can parse a wide spectrum of RNA grammar architectures (including the standard nearest-neighbor model and more) using a generalized super-grammar that can be parameterized with probabilities, energies, or arbitrary scores. By using TORNADO, we find that probabilistic nearest-neighbor models perform comparably to (but not significantly better than) discriminative methods. We find that complex statistical models are prone to overfitting RNA structure and that evaluations should use structurally nonhomologous training and test data sets. Overfitting has affected at least one published method (ContextFold). The most important barrier to improving statistical approaches for RNA secondary structure prediction is the lack of diversity of well-curated single-sequence RNA secondary structures in current RNA databases.     

Viral IRES Prediction System - a Web Server for Prediction of the IRES Secondary Structure In Silico

The internal ribosomal entry site (IRES) functions as cap-independent translation initiation sites in eukaryotic cells. IRES elements have been applied as useful tools for bi-cistronic expression vectors. Current RNA structure prediction programs are unable to predict precisely the potential IRES element. We have designed a viral IRES prediction system (VIPS) to perform the IRES secondary structure prediction. In order to obtain better results for the IRES prediction, the VIPS can evaluate and predict for all four different groups of IRESs with a higher accuracy. RNA secondary structure prediction, comparison, and pseudoknot prediction programs were implemented to form the three-stage procedure for the VIPS. The backbone of VIPS includes: the RNAL fold program, aimed to predict local RNA secondary structures by minimum free energy method; the RNA Align program, intended to compare predicted structures; and pknotsRG program, used to calculate the pseudoknot structure. VIPS was evaluated by using UTR database, IRES database and Virus database, and the accuracy rate of VIPS was assessed as 98.53%, 90.80%, 82.36% and 80.41% for IRES groups 1, 2, 3, and 4, respectively. This advance useful search approach for IRES structures will facilitate IRES related studies. The VIPS on-line website service is available at http://140.135.61.250/vips/.     

INFO-RNA--a fast approach to inverse RNA folding

MOTIVATION: The structure of RNA molecules is often crucial for their function. Therefore, secondary structure prediction has gained much interest. Here, we consider the inverse RNA folding problem, which means designing RNA sequences that fold into a given structure. RESULTS: We introduce a new algorithm for the inverse folding problem (INFO-RNA) that consists of two parts; a dynamic programming method for good initial sequences and a following improved stochastic local search that uses an effective neighbor selection method. During the initialization, we design a sequence that among all sequences adopts the given structure with the lowest possible energy. For the selection of neighbors during the search, we use a kind of look-ahead of one selection step applying an additional energy-based criterion. Afterwards, the pre-ordered neighbors are tested using the actual optimization criterion of minimizing the structure distance between the target structure and the mfe structure of the considered neighbor. We compared our algorithm to RNAinverse and RNA-SSD for artificial and biological test sets. Using INFO-RNA, we performed better than RNAinverse and in most cases, we gained better results than RNA-SSD, the probably best inverse RNA folding tool on the market. AVAILABILITY: www.bioinf.uni-freiburg.de?Subpages/software.html.     

Detecting riboSNitches with RNA folding algorithms: a genome-wide benchmark

Ribonucleic acid (RNA) secondary structure prediction continues to be a significant challenge, in particular when attempting to model sequences with less rigidly defined structures, such as messenger and non-coding RNAs. Crucial to interpreting RNA structures as they pertain to individual phenotypes is the ability to detect RNAs with large structural disparities caused by a single nucleotide variant (SNV) or riboSNitches. A recently published human genome-wide parallel analysis of RNA structure (PARS) study identified a large number of riboSNitches as well as non-riboSNitches, providing an unprecedented set of RNA sequences against which to benchmark structure prediction algorithms. Here we evaluate 11 different RNA folding algorithms’ riboSNitch prediction performance on these data. We find that recent algorithms designed specifically to predict the effects of SNVs on RNA structure, in particular remuRNA, RNAsnp and SNPfold, perform best on the most rigorously validated subsets of the benchmark data. In addition, our benchmark indicates that general structure prediction algorithms (e.g. RNAfold and RNAstructure) have overall better performance if base pairing probabilities are considered rather than minimum free energy calculations. Although overall aggregate algorithmic performance on the full set of riboSNitches is relatively low, significant improvement is possible if the highest confidence predictions are evaluated independently.     

蛋白质二级结构在线服务器预测评估

蛋白质二级结构的预测,对于研究蛋白质的功能和人类生命科学意义非凡。1951年开始提出预测蛋白质二级结构,1983年对于二级结构的预测只有50%的准确率。经过多年的发展,预测方式不断的改进和完善,到如今准确率已经超过80%。但目前预测在线服务器繁多,连续自动模型评估(CAMEO)也只给出服务器三级结构的预测评估,二级结构评估还未实现。针对上述问题,选取了以下6个服务器:PSRSM、MUFOLD、SPIDER、RAPTORX、JPRED和PSIPRED,对其预测的二级结构进行评估。并且为保证测试集不在训练集内,实验数据选取蛋白质结构数据库(Protein Data Bank,PDB)最新发布的蛋白质。在基于蛋白质同源性30%、50%和70%的实验中,PSRSM取得Q3的准确率分别为91.44%、88.12%和90.17%,比其他预测服务器中最高的MUFOLD分别高出3.19%、1.33%和2.19%,证明在同一类同源性数据中PSRSM比其他服务器有更好的预测效果。除此之外实验也得到其预测的Sov准确度也比其他服务器要高。比较各类服务器的方法与结果,得出今后蛋白质二级结构预测应当重点从大数据、模板和深度学习的角度进行研究。     

A new method to predict the consensus secondary structure of a set of unaligned RNA sequences

MOTIVATION: To predict the consensus secondary structure, possibly including pseudoknots, of a set of RNA unaligned sequences. RESULTS: We have designed a method based on a new representation of any RNA secondary structure as a set of structural relationships between the helices of the structure. We refer to this representation as a structural pattern. In a first step, we use thermodynamic parameters to select, for each sequence, the best secondary structures according to energy minimization and we represent each of them using its corresponding structural pattern. In a second step, we search for the repeated structural patterns, i.e. the largest structural patterns that occur in at least one sequence, i.e. included in at least one of the structural patterns associated to each sequence. Thanks to an efficient encoding of structural patterns, this search comes down to identifying the largest repeated word suffixes in a dictionary. In a third step, we compute the plausibility of each repeated structural pattern by checking if it occurs more frequently in the studied sequences than in random RNA sequences. We then suppose that the consensus secondary structure corresponds to the repeated structural pattern that displays the highest plausibility. We present several experiments concerning tRNA, fragments of 16S rRNA and 10Sa RNA (including pseudoknots); in each of them, we found the putative consensus secondary structure.     

Prediction of geometrically feasible three-dimensional structures of pseudoknotted RNA through free energy estimation

Accurate free energy estimation is essential for RNA structure prediction. The widely used Turner''s energy model works well for nested structures. For pseudoknotted RNAs, however, there is no effective rule for estimation of loop entropy and free energy. In this work we present a new free energy estimation method, termed the pseudoknot predictor in three-dimensional space (pk3D), which goes beyond Turner''s model. Our approach treats nested and pseudoknotted structures alike in one unifying physical framework, regardless of how complex the RNA structures are. We first test the ability of pk3D in selecting native structures from a large number of decoys for a set of 43 pseudoknotted RNA molecules, with lengths ranging from 23 to 113. We find that pk3D performs slightly better than the Dirks and Pierce extension of Turner''s rule. We then test pk3D for blind secondary structure prediction, and find that pk3D gives the best sensitivity and comparable positive predictive value (related to specificity) in predicting pseudoknotted RNA secondary structures, when compared with other methods. A unique strength of pk3D is that it also generates spatial arrangement of structural elements of the RNA molecule. Comparison of three-dimensional structures predicted by pk3D with the native structure measured by nuclear magnetic resonance or X-ray experiments shows that the predicted spatial arrangement of stems and loops is often similar to that found in the native structure. These close-to-native structures can be used as starting points for further refinement to derive accurate three-dimensional structures of RNA molecules, including those with pseudoknots.     

Statistical evaluation of improvement in RNA secondary structure prediction

With discovery of diverse roles for RNA, its centrality in cellular functions has become increasingly apparent. A number of algorithms have been developed to predict RNA secondary structure. Their performance has been benchmarked by comparing structure predictions to reference secondary structures. Generally, algorithms are compared against each other and one is selected as best without statistical testing to determine whether the improvement is significant. In this work, it is demonstrated that the prediction accuracies of methods correlate with each other over sets of sequences. One possible reason for this correlation is that many algorithms use the same underlying principles. A set of benchmarks published previously for programs that predict a structure common to three or more sequences is statistically analyzed as an example to show that it can be rigorously evaluated using paired two-sample t-tests. Finally, a pipeline of statistical analyses is proposed to guide the choice of data set size and performance assessment for benchmarks of structure prediction. The pipeline is applied using 5S rRNA sequences as an example.     

A new algorithm for RNA secondary structure design

The function of many RNAs depends crucially on their structure. Therefore, the design of RNA molecules with specific structural properties has many potential applications, e.g. in the context of investigating the function of biological RNAs, of creating new ribozymes, or of designing artificial RNA nanostructures. Here, we present a new algorithm for solving the following RNA secondary structure design problem: given a secondary structure, find an RNA sequence (if any) that is predicted to fold to that structure. Unlike the (pseudoknot-free) secondary structure prediction problem, this problem appears to be hard computationally. Our new algorithm, "RNA Secondary Structure Designer (RNA-SSD)", is based on stochastic local search, a prominent general approach for solving hard combinatorial problems. A thorough empirical evaluation on computationally predicted structures of biological sequences and artificially generated RNA structures as well as on empirically modelled structures from the biological literature shows that RNA-SSD substantially out-performs the best known algorithm for this problem, RNAinverse from the Vienna RNA Package. In particular, the new algorithm is able to solve structures, consistently, for which RNAinverse is unable to find solutions. The RNA-SSD software is publically available under the name of RNA Designer at the RNASoft website (www.rnasoft.ca).     

RAG-3D: a search tool for RNA 3D substructures

To address many challenges in RNA structure/function prediction, the characterization of RNA''s modular architectural units is required. Using the RNA-As-Graphs (RAG) database, we have previously explored the existence of secondary structure (2D) submotifs within larger RNA structures. Here we present RAG-3D—a dataset of RNA tertiary (3D) structures and substructures plus a web-based search tool—designed to exploit graph representations of RNAs for the goal of searching for similar 3D structural fragments. The objects in RAG-3D consist of 3D structures translated into 3D graphs, cataloged based on the connectivity between their secondary structure elements. Each graph is additionally described in terms of its subgraph building blocks. The RAG-3D search tool then compares a query RNA 3D structure to those in the database to obtain structurally similar structures and substructures. This comparison reveals conserved 3D RNA features and thus may suggest functional connections. Though RNA search programs based on similarity in sequence, 2D, and/or 3D structural elements are available, our graph-based search tool may be advantageous for illuminating similarities that are not obvious; using motifs rather than sequence space also reduces search times considerably. Ultimately, such substructuring could be useful for RNA 3D structure prediction, structure/function inference and inverse folding.     

A set of nearest neighbor parameters for predicting the enthalpy change of RNA secondary structure formation

A complete set of nearest neighbor parameters to predict the enthalpy change of RNA secondary structure formation was derived. These parameters can be used with available free energy nearest neighbor parameters to extend the secondary structure prediction of RNA sequences to temperatures other than 37°C. The parameters were tested by predicting the secondary structures of sequences with known secondary structure that are from organisms with known optimal growth temperatures. Compared with the previous set of enthalpy nearest neighbor parameters, the sensitivity of base pair prediction improved from 65.2 to 68.9% at optimal growth temperatures ranging from 10 to 60°C. Base pair probabilities were predicted with a partition function and the positive predictive value of structure prediction is 90.4% when considering the base pairs in the lowest free energy structure with pairing probability of 0.99 or above. Moreover, a strong correlation is found between the predicted melting temperatures of RNA sequences and the optimal growth temperatures of the host organism. This indicates that organisms that live at higher temperatures have evolved RNA sequences with higher melting temperatures.     

Protein beta-turn prediction using nearest-neighbor method

MOTIVATION: With the emerging success of protein secondary structure prediction through the applications of various statistical and machine learning techniques, similar techniques have been applied to protein beta-turn prediction. In this study, we perform protein beta-turn prediction using a k-nearest neighbor method, which is combined with a filter that uses predicted protein secondary structure information. Traditional beta-turn prediction from k-nearest neighbor method is modified to account for the unbalanced ratio of the natural occurrence of beta-turns and non-beta-turns. RESULTS: Our prediction scheme is tested on a set of 426 non-homologous protein sequences. The prediction scheme consists of two stages: k-nearest neighbor method stage and filtering stage. Variations of the k-nearest neighbor method were used to take property of beta-turns into consideration. Our filtering method uses beta-turn/non-beta-turn estimates from the k-nearest neighbor method stage and predicted protein secondary structure information from PSI-PRED in order to get new beta-turn/non-beta-turn estimate. Our result is compared with the previously best known beta-turn prediction method on the dataset of 426 non-homologous protein sequences and is shown to give slightly superior performance at significantly lower computational complexity. AVAILABILITY: Contact the author for information on the source code of the programs used.     

Novel and efficient RNA secondary structure prediction using hierarchical folding.

Algorithms for prediction of RNA secondary structure-the set of base pairs that form when an RNA molecule folds-are valuable to biologists who aim to understand RNA structure and function. Improving the accuracy and efficiency of prediction methods is an ongoing challenge, particularly for pseudoknotted secondary structures, in which base pairs overlap. This challenge is biologically important, since pseudoknotted structures play essential roles in functions of many RNA molecules, such as splicing and ribosomal frameshifting. State-of-the-art methods, which are based on free energy minimization, have high run-time complexity (typically Theta(n(5)) or worse), and can handle (minimize over) only limited types of pseudoknotted structures. We propose a new approach for prediction of pseudoknotted structures, motivated by the hypothesis that RNA structures fold hierarchically, with pseudoknot-free (non-overlapping) base pairs forming first, and pseudoknots forming later so as to minimize energy relative to the folded pseudoknot-free structure. Our HFold algorithm uses two-phase energy minimization to predict hierarchically formed secondary structures in O(n(3)) time, matching the complexity of the best algorithms for pseudoknot-free secondary structure prediction via energy minimization. Our algorithm can handle a wide range of biological structures, including kissing hairpins and nested kissing hairpins, which have previously required Theta(n(6)) time.