Combining prediction of secondary structure and solvent accessibility in proteins

Owing to the use of evolutionary information and advanced machine learning protocols, secondary structures of amino acid residues in proteins can be predicted from the primary sequence with more than 75% per-residue accuracy for the 3-state (i.e., helix, beta-strand, and coil) classification problem. In this work we investigate whether further progress may be achieved by incorporating the relative solvent accessibility (RSA) of an amino acid residue as a fingerprint of the overall topology of the protein. Toward that goal, we developed a novel method for secondary structure prediction that uses predicted RSA in addition to attributes derived from evolutionary profiles. Our general approach follows the 2-stage protocol of Rost and Sander, with a number of Elman-type recurrent neural networks (NNs) combined into a consensus predictor. The RSA is predicted using our recently developed regression-based method that provides real-valued RSA, with the overall correlation coefficients between the actual and predicted RSA of about 0.66 in rigorous tests on independent control sets. Using the predicted RSA, we were able to improve the performance of our secondary structure prediction by up to 1.4% and achieved the overall per-residue accuracy between 77.0% and 78.4% for the 3-state classification problem on different control sets comprising, together, 603 proteins without homology to proteins included in the training. The effects of including solvent accessibility depend on the quality of RSA prediction. In the limit of perfect prediction (i.e., when using the actual RSA values derived from known protein structures), the accuracy of secondary structure prediction increases by up to 4%. We also observed that projecting real-valued RSA into 2 discrete classes with the commonly used threshold of 25% RSA decreases the classification accuracy for secondary structure prediction. While the level of improvement of secondary structure prediction may be different for prediction protocols that implicitly account for RSA in other ways, we conclude that an increase in the 3-state classification accuracy may be achieved when combining RSA with a state-of-the-art protocol utilizing evolutionary profiles. The new method is available through a Web server at http://sable.cchmc.org.     

Linear regression models for solvent accessibility prediction in proteins.

The relative solvent accessibility (RSA) of an amino acid residue in a protein structure is a real number that represents the solvent exposed surface area of this residue in relative terms. The problem of predicting the RSA from the primary amino acid sequence can therefore be cast as a regression problem. Nevertheless, RSA prediction has so far typically been cast as a classification problem. Consequently, various machine learning techniques have been used within the classification framework to predict whether a given amino acid exceeds some (arbitrary) RSA threshold and would thus be predicted to be "exposed," as opposed to "buried." We have recently developed novel methods for RSA prediction using nonlinear regression techniques which provide accurate estimates of the real-valued RSA and outperform classification-based approaches with respect to commonly used two-class projections. However, while their performance seems to provide a significant improvement over previously published approaches, these Neural Network (NN) based methods are computationally expensive to train and involve several thousand parameters. In this work, we develop alternative regression models for RSA prediction which are computationally much less expensive, involve orders-of-magnitude fewer parameters, and are still competitive in terms of prediction quality. In particular, we investigate several regression models for RSA prediction using linear L1-support vector regression (SVR) approaches as well as standard linear least squares (LS) regression. Using rigorously derived validation sets of protein structures and extensive cross-validation analysis, we compare the performance of the SVR with that of LS regression and NN-based methods. In particular, we show that the flexibility of the SVR (as encoded by metaparameters such as the error insensitivity and the error penalization terms) can be very beneficial to optimize the prediction accuracy for buried residues. We conclude that the simple and computationally much more efficient linear SVR performs comparably to nonlinear models and thus can be used in order to facilitate further attempts to design more accurate RSA prediction methods, with applications to fold recognition and de novo protein structure prediction methods.     

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.     

Prediction of protein relative solvent accessibility with a two-stage SVM approach

Information on relative solvent accessibility (RSA) of amino acid residues in proteins provides valuable clues to the prediction of protein structure and function. A two-stage approach with support vector machines (SVMs) is proposed, where an SVM predictor is introduced to the output of the single-stage SVM approach to take into account the contextual relationships among solvent accessibilities for the prediction. By using the position-specific scoring matrices (PSSMs) generated by PSI-BLAST, the two-stage SVM approach achieves accuracies up to 90.4% and 90.2% on the Manesh data set of 215 protein structures and the RS126 data set of 126 nonhomologous globular proteins, respectively, which are better than the highest published scores on both data sets to date. A Web server for protein RSA prediction using a two-stage SVM method has been developed and is available (http://birc.ntu.edu.sg/~pas0186457/rsa.html).     

A hydrophobic spine stabilizes a surface-exposed α-helix according to analysis of the solvent-accessible surface area

Background

Most of hydrophilic and hydrophobic residues are thought to be exposed and buried in proteins, respectively. In contrast to the majority of the existing studies on protein folding characteristics using protein structures, in this study, our aim was to design predictors for estimating relative solvent accessibility (RSA) of amino acid residues to discover protein folding characteristics from sequences.

Methods

The proposed 20 real-value RSA predictors were designed on the basis of the support vector regression method with a set of informative physicochemical properties (PCPs) obtained by means of an optimal feature selection algorithm. Then, molecular dynamics simulations were performed for validating the knowledge discovered by analysis of the selected PCPs.

Results

The RSA predictors had the mean absolute error of 14.11% and a correlation coefficient of 0.69, better than the existing predictors. The hydrophilic-residue predictors preferred PCPs of buried amino acid residues to PCPs of exposed ones as prediction features. A hydrophobic spine composed of exposed hydrophobic residues of an α-helix was discovered by analyzing the PCPs of RSA predictors corresponding to hydrophobic residues. For example, the results of a molecular dynamics simulation of wild-type sequences and their mutants showed that proteins 1MOF and 2WRP_H16I (Protein Data Bank IDs), which have a perfectly hydrophobic spine, have more stable structures than 1MOF_I54D and 2WRP do (which do not have a perfectly hydrophobic spine).

Conclusions

We identified informative PCPs to design high-performance RSA predictors and to analyze these PCPs for identification of novel protein folding characteristics. A hydrophobic spine in a protein can help to stabilize exposed α-helices.

     

Prediction-based fingerprints of protein-protein interactions

The recognition of protein interaction sites is an important intermediate step toward identification of functionally relevant residues and understanding protein function, facilitating experimental efforts in that regard. Toward that goal, the authors propose a novel representation for the recognition of protein-protein interaction sites that integrates enhanced relative solvent accessibility (RSA) predictions with high resolution structural data. An observation that RSA predictions are biased toward the level of surface exposure consistent with protein complexes led the authors to investigate the difference between the predicted and actual (i.e., observed in an unbound structure) RSA of an amino acid residue as a fingerprint of interaction sites. The authors demonstrate that RSA prediction-based fingerprints of protein interactions significantly improve the discrimination between interacting and noninteracting sites, compared with evolutionary conservation, physicochemical characteristics, structure-derived and other features considered before. On the basis of these observations, the authors developed a new method for the prediction of protein-protein interaction sites, using machine learning approaches to combine the most informative features into the final predictor. For training and validation, the authors used several large sets of protein complexes and derived from them nonredundant representative chains, with interaction sites mapped from multiple complexes. Alternative machine learning techniques are used, including Support Vector Machines and Neural Networks, so as to evaluate the relative effects of the choice of a representation and a specific learning algorithm. The effects of induced fit and uncertainty of the negative (noninteracting) class assignment are also evaluated. Several representative methods from the literature are reimplemented to enable direct comparison of the results. Using rigorous validation protocols, the authors estimated that the new method yields the overall classification accuracy of about 74% and Matthews correlation coefficients of 0.42, as opposed to up to 70% classification accuracy and up to 0.3 Matthews correlation coefficient for methods that do not utilize RSA prediction-based fingerprints. The new method is available at http://sppider.cchmc.org.     

Understanding the Effect of Secondary Structures and Aggregation on Human Protein Folding Class Evolution

Using several model organisms it has been shown earlier that protein designability is related to contact density or fraction of buried residues and influence protein evolutionary rates dramatically. Here, using Homo sapiens as a model organism, we have analyzed two main folding classes (all-α and all-β) to examine the factors affecting their evolutionary rates. Since, secondary structures are the most fundamental components of the protein folding classes, we explored the effect of protein secondary structure composition on evolution. Our results show that sheet and helix fractions exhibit positive and negative correlations, respectively, with the rate of protein evolution. On dividing the secondary structure components according to solvent accessibility, linear regression model identified two factors namely buried sheet fraction and relative aggregation propensity. Both these factors together can explain about 13.4% variability in the rate of human protein evolution, while buried sheet residues can alone account to 9.9% variability.     

On the relation between residue flexibility and local solvent accessibility in proteins

We investigate the relationship between the flexibility, expressed with B‐factor, and the relative solvent accessibility (RSA) in the context of local, with respect to the sequence, neighborhood and related concepts such as residue depth. We observe that the flexibility of a given residue is strongly influenced by the solvent accessibility of the adjacent neighbors. The mean normalized B‐factor of the exposed residues with two buried neighbors is smaller than that of the buried residues with two exposed neighbors. Inclusion of RSA of the neighboring residues (local RSA) significantly increases correlation with the B‐factor. Correlation between the local RSA and B‐factor is shown to be stronger than the correlation that considers local distance‐ or volume‐based residue depth. We also found that the correlation coefficients between B‐factor and RSA for the 20 amino acids, called flexibility‐exposure correlation index, are strongly correlated with the stability scale that characterizes the average contributions of each amino acid to the folding stability. Our results reveal that the predicted RSA could be used to distinguish between the disordered and ordered residues and that the inclusion of local predicted RSA values helps providing a better contrast between these two types of residues. Prediction models developed based on local actual RSA and local predicted RSA show similar or better results in the context of B‐factor and disorder predictions when compared with several existing approaches. We validate our models using three case studies, which show that this work provides useful clues for deciphering the structure–flexibility–function relation. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

A hierarchical approach to all-atom protein loop prediction

The application of all-atom force fields (and explicit or implicit solvent models) to protein homology-modeling tasks such as side-chain and loop prediction remains challenging both because of the expense of the individual energy calculations and because of the difficulty of sampling the rugged all-atom energy surface. Here we address this challenge for the problem of loop prediction through the development of numerous new algorithms, with an emphasis on multiscale and hierarchical techniques. As a first step in evaluating the performance of our loop prediction algorithm, we have applied it to the problem of reconstructing loops in native structures; we also explicitly include crystal packing to provide a fair comparison with crystal structures. In brief, large numbers of loops are generated by using a dihedral angle-based buildup procedure followed by iterative cycles of clustering, side-chain optimization, and complete energy minimization of selected loop structures. We evaluate this method by using the largest test set yet used for validation of a loop prediction method, with a total of 833 loops ranging from 4 to 12 residues in length. Average/median backbone root-mean-square deviations (RMSDs) to the native structures (superimposing the body of the protein, not the loop itself) are 0.42/0.24 A for 5 residue loops, 1.00/0.44 A for 8 residue loops, and 2.47/1.83 A for 11 residue loops. Median RMSDs are substantially lower than the averages because of a small number of outliers; the causes of these failures are examined in some detail, and many can be attributed to errors in assignment of protonation states of titratable residues, omission of ligands from the simulation, and, in a few cases, probable errors in the experimentally determined structures. When these obvious problems in the data sets are filtered out, average RMSDs to the native structures improve to 0.43 A for 5 residue loops, 0.84 A for 8 residue loops, and 1.63 A for 11 residue loops. In the vast majority of cases, the method locates energy minima that are lower than or equal to that of the minimized native loop, thus indicating that sampling rarely limits prediction accuracy. The overall results are, to our knowledge, the best reported to date, and we attribute this success to the combination of an accurate all-atom energy function, efficient methods for loop buildup and side-chain optimization, and, especially for the longer loops, the hierarchical refinement protocol.     

Prediction of turn types in protein structure by machine-learning classifiers

We present machine learning approaches for turn prediction from the amino acid sequence. Different turn classes and types were considered based on a novel turn classification scheme. We trained an unsupervised (self-organizing map) and two kernel-based classifiers, namely the support vector machine and a probabilistic neural network. Turn versus non-turn classification was carried out for turn families containing intramolecular hydrogen bonds and three to six residues. Support vector machine classifiers yielded a Matthews correlation coefficient (mcc) of approximately 0.6 and a prediction accuracy of 80%. Probabilistic neural networks were developed for beta-turn type prediction. The method was able to distinguish between five types of beta-turns yielding mcc > 0.5 and at least 80% overall accuracy. We conclude that the proposed new turn classification is distinct and well-defined, and machine learning classifiers are suited for sequence-based turn prediction. Their potential for sequence-based prediction of turn structures is discussed.     

Prediction of protein solvent accessibility using fuzzy k-nearest neighbor method

MOTIVATION: The solvent accessibility of amino acid residues plays an important role in tertiary structure prediction, especially in the absence of significant sequence similarity of a query protein to those with known structures. The prediction of solvent accessibility is less accurate than secondary structure prediction in spite of improvements in recent researches. The k-nearest neighbor method, a simple but powerful classification algorithm, has never been applied to the prediction of solvent accessibility, although it has been used frequently for the classification of biological and medical data. RESULTS: We applied the fuzzy k-nearest neighbor method to the solvent accessibility prediction, using PSI-BLAST profiles as feature vectors, and achieved high prediction accuracies. With leave-one-out cross-validation on the ASTRAL SCOP reference dataset constructed by sequence clustering, our method achieved 64.1% accuracy for a 3-state (buried/intermediate/exposed) prediction (thresholds of 9% for buried/intermediate and 36% for intermediate/exposed) and 86.7, 82.0, 79.0 and 78.5% accuracies for 2-state (buried/exposed) predictions (thresholds of each 0, 5, 16 and 25% for buried/exposed), respectively. Our method also showed slightly better accuracies than other methods by about 2-5% on the RS126 dataset and a benchmarking dataset with 229 proteins. AVAILABILITY: Program and datasets are available at http://biocom1.ssu.ac.kr/FKNNacc/ CONTACT: jul@ssu.ac.kr.     

Identification of functional subclasses in the DJ-1 superfamily proteins

Genomics has posed the challenge of determination of protein function from sequence and/or 3-D structure. Functional assignment from sequence relationships can be misleading, and structural similarity does not necessarily imply functional similarity. Proteins in the DJ-1 family, many of which are of unknown function, are examples of proteins with both sequence and fold similarity that span multiple functional classes. THEMATICS (theoretical microscopic titration curves), an electrostatics-based computational approach to functional site prediction, is used to sort proteins in the DJ-1 family into different functional classes. Active site residues are predicted for the eight distinct DJ-1 proteins with available 3-D structures. Placement of the predicted residues onto a structural alignment for six of these proteins reveals three distinct types of active sites. Each type overlaps only partially with the others, with only one residue in common across all six sets of predicted residues. Human DJ-1 and YajL from Escherichia coli have very similar predicted active sites and belong to the same probable functional group. Protease I, a known cysteine protease from Pyrococcus horikoshii, and PfpI/YhbO from E. coli, a hypothetical protein of unknown function, belong to a separate class. THEMATICS predicts a set of residues that is typical of a cysteine protease for Protease I; the prediction for PfpI/YhbO bears some similarity. YDR533Cp from Saccharomyces cerevisiae, of unknown function, and the known chaperone Hsp31 from E. coli constitute a third group with nearly identical predicted active sites. While the first four proteins have predicted active sites at dimer interfaces, YDR533Cp and Hsp31 both have predicted sites contained within each subunit. Although YDR533Cp and Hsp31 form different dimers with different orientations between the subunits, the predicted active sites are superimposable within the monomer structures. Thus, the three predicted functional classes form four different types of quaternary structures. The computational prediction of the functional sites for protein structures of unknown function provides valuable clues for functional classification.     

Prediction of protein solvent accessibility using support vector machines

A Support Vector Machine learning system has been trained to predict protein solvent accessibility from the primary structure. Different kernel functions and sliding window sizes have been explored to find how they affect the prediction performance. Using a cut-off threshold of 15% that splits the dataset evenly (an equal number of exposed and buried residues), this method was able to achieve a prediction accuracy of 70.1% for single sequence input and 73.9% for multiple alignment sequence input, respectively. The prediction of three and more states of solvent accessibility was also studied and compared with other methods. The prediction accuracies are better than, or comparable to, those obtained by other methods such as neural networks, Bayesian classification, multiple linear regression, and information theory. In addition, our results further suggest that this system may be combined with other prediction methods to achieve more reliable results, and that the Support Vector Machine method is a very useful tool for biological sequence analysis.     

Two-stage support vector regression approach for predicting accessible surface areas of amino acids

We address the problem of predicting solvent accessible surface area (ASA) of amino acid residues in protein sequences, without classifying them into buried and exposed types. A two-stage support vector regression (SVR) approach is proposed to predict real values of ASA from the position-specific scoring matrices generated from PSI-BLAST profiles. By adding SVR as the second stage to capture the influences on the ASA value of a residue by those of its neighbors, the two-stage SVR approach achieves improvements of mean absolute errors up to 3.3%, and correlation coefficients of 0.66, 0.68, and 0.67 on the Manesh dataset of 215 proteins, the Barton dataset of 502 nonhomologous proteins, and the Carugo dataset of 338 proteins, respectively, which are better than the scores published earlier on these datasets. A Web server for protein ASA prediction by using a two-stage SVR method has been developed and is available (http://birc.ntu.edu.sg/~ pas0186457/asa.html).     

Accurate prediction of protein folding rates from sequence and sequence‐derived residue flexibility and solvent accessibility

Protein folding rates vary by several orders of magnitude and they depend on the topology of the fold and the size and composition of the sequence. Although recent works show that the rates can be predicted from the sequence, allowing for high‐throughput annotations, they consider only the sequence and its predicted secondary structure. We propose a novel sequence‐based predictor, PFR‐AF, which utilizes solvent accessibility and residue flexibility predicted from the sequence, to improve predictions and provide insights into the folding process. The predictor includes three linear regressions for proteins with two‐state, multistate, and unknown (mixed‐state) folding kinetics. PFR‐AF on average outperforms current methods when tested on three datasets. The proposed approach provides high‐quality predictions in the absence of similarity between the predicted and the training sequences. The PFR‐AF's predictions are characterized by high (between 0.71 and 0.95, depending on the dataset) correlation and the lowest (between 0.75 and 0.9) mean absolute errors with respect to the experimental rates, as measured using out‐of‐sample tests. Our models reveal that for the two‐state chains inclusion of solvent‐exposed Ala may accelerate the folding, while increased content of Ile may reduce the folding speed. We also demonstrate that increased flexibility of coils facilitates faster folding and that proteins with larger content of solvent‐exposed strands may fold at a slower pace. The increased flexibility of the solvent‐exposed residues is shown to elongate folding, which also holds, with a lower correlation, for buried residues. Two case studies are included to support our findings. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Ground-penetrating radar-based automatic reconstruction of three-dimensional coarse root system architecture

Background and aims

The diverse functions of roots set requirements on specific root system architecture (RSA). Investigation on RSA holds potentials for studying the adaptation of plants to environmental stresses and to interspecific competitions. Ground-penetrating radar (GPR) has provided a non-invasive method for studying in situ RSA. However, previous GPR method relied on manually connecting root points detected between radargrams to restore each root branch, resulting in limited accuracy and efficiency of reconstructing RSA. The objective of this study is to improve the effectiveness of 3D RSA reconstruction using GPR root detection data.

Methods

A total of 213 coarse root sections (with diameter >0.5 cm) were extracted from a distribution map of a reference shrub (Arctostaphylos pungens) root system to simulate the coarse roots identified by GPR. An automatic method was established to trace each root point to its optimum growing source point. Connections between discrete root points recovered the topology of the reference RSA. A spline curve smoothing method was applied to restore the 3D morphology of the reference RSA. The proposed protocol was then tested to rebuild the 3D RSA of a shrub (Caragana microphylla) growing in the sandy soils after in situ GPR survey. The accuracy of RSA reconstruction was quantitatively evaluated by a relationship matrix method and qualitatively assessed by direct comparisons between the reconstructed and the actual RSAs after in situ excavation.

Results

For both simulated and field collected GPR detection datasets, the reconstructed RSAs strongly corresponded to the real topology of the actual root systems. When adapting the best strategy, 186 of the 213 (87.32 %) root points on the reference root system of A. pungens were interlinked with correct topology, and the relationship matrix method detected an overall similarity of 82.75 % between the reconstructed and the actual RSAs.

Conclusion

The proposed automatic RSA reconstruction method greatly enhances the interpretation of GPR detection data regarding coarse roots, making in situ non-invasive and long-term mapping and monitoring of RSA possible.     

Real value prediction of solvent accessibility from amino acid sequence

The solvent accessibility of amino acid residues has been predicted in the past by classifying them into exposure states with varying thresholds. This classification provides a wide range of values for the accessible surface area (ASA) within which a residue may fall. Thus far, no attempt has been made to predict real values of ASA from the sequence information without a priori classification into exposure states. Here, we present a new method with which to predict real value ASAs for residues, based on neighborhood information. Our real value prediction neural network could estimate the ASA for four different nonhomologous, nonredundant data sets of varying size, with 18.0-19.5% mean absolute error, defined as per residue absolute difference between the predicted and experimental values of relative ASA. Correlation between the predicted and experimental values ranged from 0.47 to 0.50. It was observed that the ASA of a residue could be predicted within a 23.7% mean absolute error, even when no information about its neighbors is included. Prediction of real values answers the issue of arbitrary choice of ASA state thresholds, and carries more information than category prediction. Prediction error for each residue type strongly correlates with the variability in its experimental ASA values.     

基于支持向量机方法的蛋白可溶性预测

按照蛋白质序列中残基的相对可溶性,将其分为两类(表面／内部)和三类(表面／中间／内部)进行预测。选择不同窗宽和参数对数据进行训练和预测,以确保得到最好的分类效果,并同其他已有方法进行比较。对同一数据集不同分类阈值的预测结果显示,支持向量机方法对蛋白质可溶性的整体预测效果优于神经网络和信息论的方法。其中,对两类数据的最优分类结果达到79．0％,对三类数据的最优分类结果达到67．5％,表明支持向量机是蛋白质残基可溶性预测的一种有效方法。     

Average assignment method for predicting the stability of protein mutants

Prediction of protein stability upon amino acid substitutions is an important problem in molecular biology and it will be helpful for designing stable mutants. In this work, we have analyzed the stability of protein mutants using three different data sets of 1791, 1396, and 2204 mutants, respectively, for thermal stability (DeltaTm), free energy change due to thermal (DeltaDeltaG), and denaturant denaturations (DeltaDeltaGH2O), obtained from the ProTherm database. We have classified the mutants into 380 possible substitutions and assigned the stability of each mutant using the information obtained with similar type of mutations. We observed that this assignment could distinguish the stabilizing and destabilizing mutants to an accuracy of 70-80% at different measures of stability. Further, we have classified the mutants based on secondary structure and solvent accessibility (ASA) and observed that the classification significantly improved the accuracy of prediction. The classification of mutants based on helix, strand, and coil distinguished the stabilizing/destabilizing mutants at an average accuracy of 82% and the correlation is 0.56; information about the location of residues at the interior, partially buried, and surface regions of a protein correctly identified the stabilizing/destabilizing residues at an average accuracy of 81% and the correlation is 0.59. The nine subclassifications based on three secondary structures and solvent accessibilities improved the accuracy of assigning stabilizing/destabilizing mutants to an accuracy of 84-89% for the three data sets. Further, the present method is able to predict the free energy change (DeltaDeltaG) upon mutations within a deviation of 0.64 kcal/mol. We suggest that this method could be used for predicting the stability of protein mutants.