RNA Secondary Structure Prediction Methods Research Articles

RNAknot2S: a New RNA Secondary Structure Prediction Method that Combines the Comparative Approach with Genetic Algorithms

RNAknot2S: a New RNA Secondary Structure Prediction Method that Combines the Comparative Approach with Genetic Algorithms

RNA independent fragment partition method based on deep learning for RNA secondary structure prediction

The non-coding RNA secondary structure largely determines its function. Hence, accuracy in structure acquisition is of great importance. Currently, this acquisition primarily relies on various computational methods. The prediction of the structures of long RNA sequences with high precision and reasonable computational cost remains challenging. Here, we propose a deep learning model, RNA-par, which could partition an RNA sequence into several independent fragments (i-fragments) based on its exterior loops. Each i-fragment secondary structure predicted individually could be further assembled to acquire the complete RNA secondary structure. In the examination of our independent test set, the average length of the predicted i-fragments was 453 nt, which was considerably shorter than that of complete RNA sequences (848 nt). The accuracy of the assembled structures was higher than that of the structures predicted directly using the state-of-the-art RNA secondary structure prediction methods. This proposed model could serve as a preprocessing step for RNA secondary structure prediction for enhancing the predictive performance (especially for long RNA sequences) and reducing the computational cost. In the future, predicting the secondary structure of long-sequence RNA with high accuracy can be enabled by developing a framework combining RNA-par with various existing RNA secondary structure prediction algorithms. Our models, test codes and test data are provided at https://github.com/mianfei71/RNAPar.

Review of machine learning methods for RNA secondary structure prediction.

Secondary structure plays an important role in determining the function of noncoding RNAs. Hence, identifying RNA secondary structures is of great value to research. Computational prediction is a mainstream approach for predicting RNA secondary structure. Unfortunately, even though new methods have been proposed over the past 40 years, the performance of computational prediction methods has stagnated in the last decade. Recently, with the increasing availability of RNA structure data, new methods based on machine learning (ML) technologies, especially deep learning, have alleviated the issue. In this review, we provide a comprehensive overview of RNA secondary structure prediction methods based on ML technologies and a tabularized summary of the most important methods in this field. The current pending challenges in the field of RNA secondary structure prediction and future trends are also discussed.

A New Method of RNA Secondary Structure Prediction Based on Convolutional Neural Network and Dynamic Programming.

In recent years, obtaining RNA secondary structure information has played an important role in RNA and gene function research. Although some RNA secondary structures can be gained experimentally, in most cases, efficient, and accurate computational methods are still needed to predict RNA secondary structure. Current RNA secondary structure prediction methods are mainly based on the minimum free energy algorithm, which finds the optimal folding state of RNA in vivo using an iterative method to meet the minimum energy or other constraints. However, due to the complexity of biotic environment, a true RNA structure always keeps the balance of biological potential energy status, rather than the optimal folding status that meets the minimum energy. For short sequence RNA its equilibrium energy status for the RNA folding organism is close to the minimum free energy status; therefore, the minimum free energy algorithm for predicting RNA secondary structure has higher accuracy. Nevertheless, in a longer sequence RNA, constant folding causes its biopotential energy balance to deviate far from the minimum free energy status. This deviation is because of its complex structure and results in a serious decline in the prediction accuracy of its secondary structure. In this paper, we propose a novel RNA secondary structure prediction algorithm using a convolutional neural network model combined with a dynamic programming method to improve the accuracy with large-scale RNA sequence and structure data. We analyze current experimental RNA sequences and structure data to construct a deep convolutional network model, and then we extract implicit features of an effective classification from large-scale data to predict the pairing probability of each base in an RNA sequence. For the obtained probabilities of RNA sequence base pairing, an enhanced dynamic programming method is applied to obtain the optimal RNA secondary structure. Results indicate that our proposed method is superior to the common RNA secondary structure prediction algorithms in predicting three benchmark RNA families. Based on the characteristics of deep learning algorithm, it can be inferred that the method proposed in this paper has a 30% higher prediction success rate when compared with other algorithms, which will be needed as the amount of real RNA structure data increases in the future.

Towards an understanding of RNA structural modalities: a riboswitch case study.

A riboswitch is a type of RNA molecule that regulates important biological functions by changing structure, typically under ligand-binding. We assess the extent that these ligand-bound structural alternatives are present in the Boltzmann sample, a standard RNA secondary structure prediction method, for three riboswitch test cases. We use the cluster analysis tool RNAStructProfiling to characterize the different modalities present among the suboptimal structures sampled. We compare these modalities to the putative base pairing models obtained from independent experiments using NMR or fluorescence spectroscopy. We find, somewhat unexpectedly, that profiling the Boltzmann sample captures evidence of ligand-bound conformations for two of three riboswitches studied. Moreover, this agreement between predicted modalities and experimental models is consistent with the classification of riboswitches into thermodynamic versus kinetic regulatory mechanisms. Our results support cluster analysis of Boltzmann samples by RNAStructProfiling as a possible basis for de novo identification of thermodynamic riboswitches, while highlighting the challenges for kinetic ones.

Consensus RNA secondary structure prediction using information of neighbouring columns and principal component analysis

RNA is a family of biological macromolecules. It is important to all kinds of biological processes. RNA structure is closely related to its functions. Hence, determining the structure is invaluable in understanding genetic diseases and creating drugs. Nowadays, RNA secondary structure prediction is a field yet to be researched. In this paper, we present a novel method using RNA sequence alignment to predict a consensus RNA secondary structure. In essence, the goal of the method is to give a prediction about whether any two columns of an alignment correspond to a base pair or not, using the information provided by the alignment. The information includes the covariation score, the fraction of complementary nucleotides and the consensus probability matrix of the column pair and those of its neighbours. Then principal component analysis is applied to overcome the problem of over-fitting. A comparison of our method and other consensus RNA secondary structure prediction methods including NeCFold, ELMFold, KnetFold, PFold and RNAalifold, in 47 families from Rfam (version 11.0) is performed. Results show that our method surpasses the other methods in terms of Matthews correlation coefficient, sensitivity and selectivity.

Evaluation of RNA secondary structure prediction for both base-pairing and topology

Graphical abstractSecondary structures of RNAs are crucial to the understanding of their tertiary structures and functions. At present, many theoretical methods are widely used to predict RNA secondary structures. The performance of these methods has been evaluated but only for their ability of base-pairing prediction. However, the topology of a RNA secondary structure is more important for understanding its tertiary structure and function, especially for long RNAs. In this paper, we constructed a new non-redundant RNA database containing 73 RNA with lengths of 50–300 nucleotides and benchmarked four popular algorithms for both base pairing and topology. The results show that the prediction accuracy of secondary structure topology is only 38%, in contrast to 70% for that of base pairing. Furthermore, the topological consistency is not strongly correlated to the base-pairing consistency. Our results will be helpful to understand the limitations of RNA secondary structure prediction methods from a different point of view and also to their improvements in future.

Conditioning and Robustness of RNA Boltzmann Sampling under Thermodynamic Parameter Perturbations

Understanding how RNA secondary structure prediction methods depend on the underlying nearest-neighbor thermodynamic model remains a fundamental challenge in the field. Minimum free energy (MFE) predictions are known to be “ill conditioned” in that small changes to the thermodynamic model can result in significantly different optimal structures. Hence, the best practice is now to sample from the Boltzmann distribution, which generates a set of suboptimal structures. Although the structural signal of this Boltzmann sample is known to be robust to stochastic noise, the conditioning and robustness under thermodynamic perturbations have yet to be addressed. We present here a mathematically rigorous model for conditioning inspired by numerical analysis, and also a biologically inspired definition for robustness under thermodynamic perturbation. We demonstrate the strong correlation between conditioning and robustness and use its tight relationship to define quantitative thresholds for well versus ill conditioning. These resulting thresholds demonstrate that the majority of the sequences are at least sample robust, which verifies the assumption of sampling’s improved conditioning over the MFE prediction. Furthermore, because we find no correlation between conditioning and MFE accuracy, the presence of both well- and ill-conditioned sequences indicates the continued need for both thermodynamic model refinements and alternate RNA structure prediction methods beyond the physics-based ones.

A New Method to Predict RNA Secondary Structure Based on RNA Folding Simulation.

RNA plays an important role in various biological processes; hence, it is essential when determining the functions of RNA to research its secondary structures. So far, the accuracy of RNA secondary structure prediction remains an area in need of improvement. This paper presents a novel method for predicting RNA secondary structure based on an RNA folding simulation model. This model assumes that the process of RNA folding from the random coil state to full structure is staged and in every stage of folding, the final state of an RNA is determined by the optimal combination of helical regions, which are urgently essential to dynamics of RNA formation. This paper proposes the First Large Free Energy Difference (FLED) in order to find the helical regions most urgently needed for optimal final state formation among all the possible helical regions. Tests on the datasets with known structures from public databases demonstrate that our method can outperform other current RNA secondary structure prediction methods in terms of prediction accuracy.

An RNA secondary structure prediction method based on minimum and suboptimal free energy structures

The function of an RNA-molecule is mainly determined by its tertiary structures. And its secondary structure is an important determinant of its tertiary structure. The comparative methods usually give better results than the single-sequence methods. Based on minimum and suboptimal free energy structures, the paper presents a novel method for predicting conserved secondary structure of a group of related RNAs. In the method, the information from the known RNA structures is used as training data in a SVM (Support Vector Machine) classifier. Our method has been tested on the benchmark dataset given by Puton et al. The results show that the average sensitivity of our method is higher than that of other comparative methods such as CentroidAlifold, MXScrana, RNAalifold, and TurboFold.

Ensemble-based prediction of RNA secondary structures.

BackgroundAccurate structure prediction methods play an important role for the understanding of RNA function. Energy-based, pseudoknot-free secondary structure prediction is one of the most widely used and versatile approaches, and improved methods for this task have received much attention over the past five years. Despite the impressive progress that as been achieved in this area, existing evaluations of the prediction accuracy achieved by various algorithms do not provide a comprehensive, statistically sound assessment. Furthermore, while there is increasing evidence that no prediction algorithm consistently outperforms all others, no work has been done to exploit the complementary strengths of multiple approaches.ResultsIn this work, we present two contributions to the area of RNA secondary structure prediction. Firstly, we use state-of-the-art, resampling-based statistical methods together with a previously published and increasingly widely used dataset of high-quality RNA structures to conduct a comprehensive evaluation of existing RNA secondary structure prediction procedures. The results from this evaluation clarify the performance relationship between ten well-known existing energy-based pseudoknot-free RNA secondary structure prediction methods and clearly demonstrate the progress that has been achieved in recent years. Secondly, we introduce AveRNA, a generic and powerful method for combining a set of existing secondary structure prediction procedures into an ensemble-based method that achieves significantly higher prediction accuracies than obtained from any of its component procedures.ConclusionsOur new, ensemble-based method, AveRNA, improves the state of the art for energy-based, pseudoknot-free RNA secondary structure prediction by exploiting the complementary strengths of multiple existing prediction procedures, as demonstrated using a state-of-the-art statistical resampling approach. In addition, AveRNA allows an intuitive and effective control of the trade-off between false negative and false positive base pair predictions. Finally, AveRNA can make use of arbitrary sets of secondary structure prediction procedures and can therefore be used to leverage improvements in prediction accuracy offered by algorithms and energy models developed in the future. Our data, MATLAB software and a web-based version of AveRNA are publicly available at http://www.cs.ubc.ca/labs/beta/Software/AveRNA.

C o F old : an RNA secondary structure prediction method that takes co-transcriptional folding into account

Existing state-of-the-art methods that take a single RNA sequence and predict the corresponding RNA secondary structure are thermodynamic methods. These aim to predict the most stable RNA structure. There exists by now ample experimental and theoretical evidence that the process of structure formation matters and that sequences in vivo fold while they are being transcribed. None of the thermodynamic methods, however, consider the process of structure formation. Here, we present a conceptually new method for predicting RNA secondary structure, called CoFold, that takes effects of co-transcriptional folding explicitly into account. Our method significantly improves the state-of-art in terms of prediction accuracy, especially for long sequences of >1000 nt in length.

CompaRNA: a server for continuous benchmarking of automated methods for RNA secondary structure prediction

We present a continuous benchmarking approach for the assessment of RNA secondary structure prediction methods implemented in the CompaRNA web server. As of 3 October 2012, the performance of 28 single-sequence and 13 comparative methods has been evaluated on RNA sequences/structures released weekly by the Protein Data Bank. We also provide a static benchmark generated on RNA 2D structures derived from the RNAstrand database. Benchmarks on both data sets offer insight into the relative performance of RNA secondary structure prediction methods on RNAs of different size and with respect to different types of structure. According to our tests, on the average, the most accurate predictions obtained by a comparative approach are generated by CentroidAlifold, MXScarna, RNAalifold and TurboFold. On the average, the most accurate predictions obtained by single-sequence analyses are generated by CentroidFold, ContextFold and IPknot. The best comparative methods typically outperform the best single-sequence methods if an alignment of homologous RNA sequences is available. This article presents the results of our benchmarks as of 3 October 2012, whereas the rankings presented online are continuously updated. We will gladly include new prediction methods and new measures of accuracy in the new editions of CompaRNA benchmarks.

A comparative taxonomy of parallel algorithms for RNA secondary structure prediction.

RNA molecules have been discovered playing crucial roles in numerous biological and medical procedures and processes. RNA structures determination have become a major problem in the biology context. Recently, computer scientists have empowered the biologists with RNA secondary structures that ease an understanding of the RNA functions and roles. Detecting RNA secondary structure is an NP-hard problem, especially in pseudoknotted RNA structures. The detection process is also time-consuming; as a result, an alternative approach such as using parallel architectures is a desirable option. The main goal in this paper is to do an intensive investigation of parallel methods used in the literature to solve the demanding issues, related to the RNA secondary structure prediction methods. Then, we introduce a new taxonomy for the parallel RNA folding methods. Based on this proposed taxonomy, a systematic and scientific comparison is performed among these existing methods.

Co-transcriptional folding is encoded within RNA genes

BackgroundMost of the existing RNA structure prediction programs fold a completely synthesized RNA molecule. However, within the cell, RNA molecules emerge sequentially during the directed process of transcription. Dedicated experiments with individual RNA molecules have shown that RNA folds while it is being transcribed and that its correct folding can also depend on the proper speed of transcription.MethodsThe main aim of this work is to study if and how co-transcriptional folding is encoded within the primary and secondary structure of RNA genes. In order to achieve this, we study the known primary and secondary structures of a comprehensive data set of 361 RNA genes as well as a set of 48 RNA sequences that are known to differ from the originally transcribed sequence units. We detect co-transcriptional folding by defining two measures of directedness which quantify the extend of asymmetry between alternative helices that lie 5' and those that lie 3' of the known helices with which they compete.ResultsWe show with statistical significance that co-transcriptional folding strongly influences RNA sequences in two ways: (1) alternative helices that would compete with the formation of the functional structure during co-transcriptional folding are suppressed and (2) the formation of transient structures which may serve as guidelines for the co-transcriptional folding pathway is encouraged.ConclusionsThese findings have a number of implications for RNA secondary structure prediction methods and the detection of RNA genes.

RNA secondary structure prediction using highly parallel computers.

An RNA secondary structure prediction method using a highly parallel computer is reported. We focus on finding thermodynamically stable structures of a single-stranded RNA molecule. Our approach is based on a parallel combinatorial method which calculates the free energy of a molecule as the sum of the free energies of all the physically possible hydrogen bonds. Our parallel algorithm finds many highly stable structures all at once, while most of the conventional prediction methods find only the most stable structure. The important idea in our algorithm is search tree pruning, with dynamic load balancing across the processor elements in a parallel computer. Software tools for visualization and classification of secondary structures are also presented using the sequence of cadang-cadang coconut viroid as an example. Our software system runs on CM-5.