Assessment of novel fold targets in CASP4: predictions of three-dimensional structures, secondary structures, and interresidue contacts

Overview Update: Computational Prediction of Intrinsic Disorder in Proteins

There are over 100 computational predictors of intrinsic disorder. These methods predict amino acid-level propensities for disorder directly from protein sequences. The propensities can be used to annotate putative disordered residues and regions. This unit provides a practical and holistic introduction to the sequence-based intrinsic disorder prediction. We define intrinsic disorder, explain the format of computational prediction of disorder, and identify and describe several accurate predictors. We also introduce recently released databases of intrinsic disorder predictions and use an illustrative example to provide insights into how predictions should be interpreted and combined. Lastly, we summarize key experimental methods that can be used to validate computational predictions. © 2023 Wiley Periodicals LLC.

Assessing the accuracy of contact and distance predictions in CASP14

We present the results of the assessment of the intramolecular residue-residue contact and distance predictions from groups participating in the 14th round of the CASP experiment. The performance of contact prediction methods was evaluated with the measures used in previous CASPs, while distance predictions were assessed based on a new protocol, which considers individual distance pairs as well as the whole predicted distance matrix, using a graph-based framework. The results of the evaluation indicate that predictions by the tFold framework, TripletRes and DeepPotential were the most accurate in both categories. With regards to progress in method performance, the results of the assessment in contact prediction did not reveal any discernible difference when compared to CASP13. Arguably, this could be due to CASP14 FM targets being more challenging than ever before.

Assessing the accuracy of contact predictions in CASP13

The accuracy of sequence-based tertiary contact predictions was assessed in a blind prediction experiment at the CASP13 meeting. After 4 years of significant improvements in prediction accuracy, another dramatic advance has taken place since CASP12 was held 2 years ago. The precision of predicting the top L/5 contacts in the free modeling category, where L is the corresponding length of the protein in residues, has exceeded 70%. As a comparison, the best-performing group at CASP12 with a 47% precision would have finished below the top 1/3 of the CASP13 groups. Extensively trained deep neural network approaches dominate the top performing algorithms, which appear to efficiently integrate information on coevolving residues and interacting fragments or possibly utilize memories of sequence similarities and sometimes can deliver accurate results even in the absence of virtually any target specific evolutionary information. If the current performance is evaluated by F-score on L contacts, it stands around 24% right now, which, despite the tremendous impact and advance in improving its utility for structure modeling, also suggests that there is much room left for further improvement.

Assessment of contact predictions in CASP12: Co-evolution and deep learning coming of age

Following up on the encouraging results of residue-residue contact prediction in the CASP11 experiment, we present the analysis of predictions submitted for CASP12. The submissions include predictions of 34 groups for 38 domains classified as free modeling targets which are not accessible to homology-based modeling due to a lack of structural templates. CASP11 saw a rise of coevolution-based methods outperforming other approaches. The improvement of these methods coupled to machine learning and sequence database growth are most likely the main driver for a significant improvement in average precision from 27% in CASP11 to 47% in CASP12. In more than half of the targets, especially those with many homologous sequences accessible, precisions above 90% were achieved with the best predictors reaching a precision of 100% in some cases. We furthermore tested the impact of using these contacts as restraints in ab initio modeling of 14 single-domain free modeling targets using Rosetta. Adding contacts to the Rosetta calculations resulted in improvements of up to 26% in GDT_TS within the top five structures.

New encouraging developments in contact prediction: Assessment of the CASP11 results

This article provides a report on the state-of-the-art in the prediction of intra-molecular residue-residue contacts in proteins based on the assessment of the predictions submitted to the CASP11 experiment. The assessment emphasis is placed on the accuracy in predicting long-range contacts. Twenty-nine groups participated in contact prediction in CASP11. At least eight of them used the recently developed evolutionary coupling techniques, with the top group (CONSIP2) reaching precision of 27% on target proteins that could not be modeled by homology. This result indicates a breakthrough in the development of methods based on the correlated mutation approach. Successful prediction of contacts was shown to be practically helpful in modeling three-dimensional structures; in particular target T0806 was modeled exceedingly well with accuracy not yet seen for ab initio targets of this size (>250 residues). Proteins 2016; 84(Suppl 1):131-144. © 2015 Wiley Periodicals, Inc.

Assessment of template-based modeling of protein structure in CASP11

We present the assessment of predictions submitted in the template-based modeling (TBM) category of CASP11 (Critical Assessment of Protein Structure Prediction). Model quality was judged on the basis of global and local measures of accuracy on all atoms including side chains. The top groups on 39 human-server targets based on model 1 predictions were LEER, Zhang, LEE, MULTICOM, and Zhang-Server. The top groups on 81 targets by server groups based on model 1 predictions were Zhang-Server, nns, BAKER-ROSETTASERVER, QUARK, and myprotein-me. In CASP11, the best models for most targets were equal to or better than the best template available in the Protein Data Bank, even for targets with poor templates. The overall performance in CASP11 is similar to the performance of predictors in CASP10 with slightly better performance on the hardest targets. For most targets, assessment measures exhibited bimodal probability density distributions. Multi-dimensional scaling of an RMSD matrix for each target typically revealed a single cluster with models similar to the target structure, with a mode in the GDT-TS density between 40 and 90, and a wide distribution of models highly divergent from each other and from the experimental structure, with density mode at a GDT-TS value of ∼20. The models in this peak in the density were either compact models with entirely the wrong fold, or highly non-compact models. The results argue for a density-driven approach in future CASP TBM assessments that accounts for the bimodal nature of these distributions instead of Z scores, which assume a unimodal, Gaussian distribution. Proteins 2016; 84(Suppl 1):200-220. © 2016 Wiley Periodicals, Inc.

JPred4: a protein secondary structure prediction server

JPred4 (http://www.compbio.dundee.ac.uk/jpred4) is the latest version of the popular JPred protein secondary structure prediction server which provides predictions by the JNet algorithm, one of the most accurate methods for secondary structure prediction. In addition to protein secondary structure, JPred also makes predictions of solvent accessibility and coiled-coil regions. The JPred service runs up to 94 000 jobs per month and has carried out over 1.5 million predictions in total for users in 179 countries. The JPred4 web server has been re-implemented in the Bootstrap framework and JavaScript to improve its design, usability and accessibility from mobile devices. JPred4 features higher accuracy, with a blind three-state (α-helix, β-strand and coil) secondary structure prediction accuracy of 82.0% while solvent accessibility prediction accuracy has been raised to 90% for residues <5% accessible. Reporting of results is enhanced both on the website and through the optional email summaries and batch submission results. Predictions are now presented in SVG format with options to view full multiple sequence alignments with and without gaps and insertions. Finally, the help-pages have been updated and tool-tips added as well as step-by-step tutorials.

A Deep Learning Network Approach to ab initio Protein Secondary Structure Prediction

Ab initio protein secondary structure (SS) predictions are utilized to generate tertiary structure predictions, which are increasingly demanded due to the rapid discovery of proteins. Although recent developments have slightly exceeded previous methods of SS prediction, accuracy has stagnated around 80 percent and many wonder if prediction cannot be advanced beyond this ceiling. Disciplines that have traditionally employed neural networks are experimenting with novel deep learning techniques in attempts to stimulate progress. Since neural networks have historically played an important role in SS prediction, we wanted to determine whether deep learning could contribute to the advancement of this field as well. We developed an SS predictor that makes use of the position-specific scoring matrix generated by PSI-BLAST and deep learning network architectures, which we call DNSS. Graphical processing units and CUDA software optimize the deep network architecture and efficiently train the deep networks. Optimal parameters for the training process were determined, and a workflow comprising three separately trained deep networks was constructed in order to make refined predictions. This deep learning network approach was used to predict SS for a fully independent test dataset of 198 proteins, achieving a Q3 accuracy of 80.7 percent and a Sov accuracy of 74.2 percent.

Context-based features enhance protein secondary structure prediction accuracy

We report a new approach of using statistical context-based scores as encoded features to train neural networks to achieve secondary structure prediction accuracy improvement. The context-based scores are pseudo-potentials derived by evaluating statistical, high-order inter-residue interactions, which estimate the favorability of a residue adopting certain secondary structure conformation within its amino acid environment. Encoding these context-based scores as important training and prediction features provides a way to address a long-standing difficulty in neural network-based secondary structure predictions of taking interdependency among secondary structures of neighboring residues into account. Our computational results have shown that the context-based scores are effective features to enhance the prediction accuracy of secondary structure predictions. An overall 7-fold cross-validated Q3 accuracy of 82.74% and Segment Overlap Accuracy (SOV) accuracy of 86.25% are achieved on a set of more than 7987 protein chains with, at most, 25% sequence identity. The Q3 prediction accuracy on benchmarks of CB513, Manesh215, Carugo338, as well as CASP9 protein chains is higher than popularly used secondary structure prediction servers, including Psipred, Profphd, Jpred, Porter (ab initio), and Netsurf. More significant improvement is observed in the SOV accuracy, where more than 4% enhancement is observed, compared to the server with the best SOV accuracy. A Q8 accuracy of >70% (71.5%) is also found in eight-state secondary structure prediction. The majority of the Q3 accuracy improvement is contributed from correctly identifying β-sheets and α-helices. When the context-based scores are incorporated, there are 15.5% more residues predicted with >90% confidence. These high-confidence predictions usually have a rather high accuracy (averagely ~95%). The three- and eight-state prediction servers (SCORPION) implementing our methods are available online.

Ab initio protein structure assembly using continuous structure fragments and optimized knowledge-based force field

Ab initio protein folding is one of the major unsolved problems in computational biology owing to the difficulties in force field design and conformational search. We developed a novel program, QUARK, for template-free protein structure prediction. Query sequences are first broken into fragments of 1-20 residues where multiple fragment structures are retrieved at each position from unrelated experimental structures. Full-length structure models are then assembled from fragments using replica-exchange Monte Carlo simulations, which are guided by a composite knowledge-based force field. A number of novel energy terms and Monte Carlo movements are introduced and the particular contributions to enhancing the efficiency of both force field and search engine are analyzed in detail. QUARK prediction procedure is depicted and tested on the structure modeling of 145 nonhomologous proteins. Although no global templates are used and all fragments from experimental structures with template modeling score >0.5 are excluded, QUARK can successfully construct 3D models of correct folds in one-third cases of short proteins up to 100 residues. In the ninth community-wide Critical Assessment of protein Structure Prediction experiment, QUARK server outperformed the second and third best servers by 18 and 47% based on the cumulative Z-score of global distance test-total scores in the FM category. Although ab initio protein folding remains a significant challenge, these data demonstrate new progress toward the solution of the most important problem in the field.

CONCORD: a consensus method for protein secondary structure prediction via mixed integer linear optimization

Most of the protein structure prediction methods use a multi-step process, which often includes secondary structure prediction, contact prediction, fragment generation, clustering, etc. For many years, secondary structure prediction has been the workhorse for numerous methods aimed at predicting protein structure and function. This paper presents a new mixed integer linear optimization (MILP)-based consensus method: a Consensus scheme based On a mixed integer liNear optimization method for seCOndary stRucture preDiction (CONCORD). Based on seven secondary structure prediction methods, SSpro, DSC, PROF, PROFphd, PSIPRED, Predator and GorIV, the MILP-based consensus method combines the strengths of different methods, maximizes the number of correctly predicted amino acids and achieves a better prediction accuracy. The method is shown to perform well compared with the seven individual methods when tested on the PDBselect25 training protein set using sixfold cross validation. It also performs well compared with another set of 10 online secondary structure prediction servers (including several recent ones) when tested on the CASP9 targets ( http://predictioncenter.org/casp9/ ). The average Q3 prediction accuracy is 83.04 per cent for the sixfold cross validation of the PDBselect25 set and 82.3 per cent for the CASP9 targets. We have developed a MILP-based consensus method for protein secondary structure prediction. A web server, CONCORD, is available to the scientific community at http://helios.princeton.edu/CONCORD .

Protein loop modeling by using fragment assembly and analytical loop closure

Protein loops are often involved in important biological functions such as molecular recognition, signal transduction, or enzymatic action. The three dimensional structures of loops can provide essential information for understanding molecular mechanisms behind protein functions. In this article, we develop a novel method for protein loop modeling, where the loop conformations are generated by fragment assembly and analytical loop closure. The fragment assembly method reduces the conformational space drastically, and the analytical loop closure method finds the geometrically consistent loop conformations efficiently. We also derive an analytic formula for the gradient of any analytical function of dihedral angles in the space of closed loops. The gradient can be used to optimize various restraints derived from experiments or databases, for example restraints for preferential interactions between specific residues or for preferred backbone angles. We demonstrate that the current loop modeling method outperforms previous methods that employ residue-based torsion angle maps or different loop closure strategies when tested on two sets of loop targets of lengths ranging from 4 to 12.

A gradient-directed Monte Carlo approach for protein design

We develop a new global optimization strategy, gradient-directed Monte Carlo (GDMC) sampling, to optimize protein sequence for a target structure using RosettaDesign. GDMC significantly improves the sampling of sequence space, compared to the classical Monte Carlo search protocol, for a fixed backbone conformation as well as for the simultaneous optimization of sequence and structure. As such, GDMC sampling enhances the efficiency of protein design.

Predicted residue-residue contacts can help the scoring of 3D models

During the 7th Critical Assessment of Protein Structure Prediction (CASP7) experiment, it was suggested that the real value of predicted residue-residue contacts might lie in the scoring of 3D model structures. Here, we have carried out a detailed reassessment of the contact predictions made during the recent CASP8 experiment to determine whether predicted contacts might aid in the selection of close-to-native structures or be a useful tool for scoring 3D structural models. We used the contacts predicted by the CASP8 residue-residue contact prediction groups to select models for each target domain submitted to the experiment. We found that the information contained in the predicted residue-residue contacts would probably have helped in the selection of 3D models in the free modeling regime and over the harder comparative modeling targets. Indeed, in many cases, the models selected using just the predicted contacts had better GDT-TS scores than all but the best 3D prediction groups. Despite the well-known low accuracy of residue-residue contact predictions, it is clear that the predictive power of contacts can be useful in 3D model prediction strategies.

Prediction of protein long-range contacts using an ensemble of genetic algorithm classifiers with sequence profile centers

Background

Prediction of long-range inter-residue contacts is an important topic in bioinformatics research. It is helpful for determining protein structures, understanding protein foldings, and therefore advancing the annotation of protein functions.

Results

In this paper, we propose a novel ensemble of genetic algorithm classifiers (GaCs) to address the long-range contact prediction problem. Our method is based on the key idea called sequence profile centers (SPCs). Each SPC is the average sequence profiles of residue pairs belonging to the same contact class or non-contact class. GaCs train on multiple but different pairs of long-range contact data (positive data) and long-range non-contact data (negative data). The negative data sets, having roughly the same sizes as the positive ones, are constructed by random sampling over the original imbalanced negative data. As a result, about 21.5% long-range contacts are correctly predicted. We also found that the ensemble of GaCs indeed makes an accuracy improvement by around 5.6% over the single GaC.

Conclusions

Classifiers with the use of sequence profile centers may advance the long-range contact prediction. In line with this approach, key structural features in proteins would be determined with high efficiency and accuracy.

A structure-centric view of protein evolution, design, and adaptation

Proteins, by virtue of their central role in most biological processes, represent one of the key subjects of the study of molecular evolution. Inherent in the indispensability of proteins for living cells is the fact that a given protein can adopt a specific three-dimensional shape that is specified solely by the protein's sequence of amino acids. Over the past several decades, structural biologists have demonstrated that the array of structures that proteins may adopt is quite astounding, and this has lead to a strong interest in understanding how protein structures change and evolve over time. In this review we consider a large body of recent work that attempts to illuminate this structure-centric picture of protein evolution. Much of this work has focused on the question of how completely new protein structures (i.e., new folds or topologies) are discovered by protein sequences as they evolve. Pursuant to this question of structural innovation has been a desire to describe and understand the observation that certain types of protein structures are far more abundant than others and how this uneven distribution of proteins implicates on the process through which new shapes are discovered. We consider a number of theoretical models that have been successful at explaining this heterogeneity in protein populations and discuss the increasing amount of evidence that indicates that the process of structural evolution involves the divergence of protein sequences and structures from one another. We also consider the topic of protein designability, which concerns itself with understanding how a protein's structure influences the number of sequences that can fold successfully into that structure. Understanding and quantifying the relationship between the physical feature of a structure and its designability has been a long-standing goal of the study of protein structure and evolution, and we discuss a number of recent advances that have yielded a promising answer to this question. Finally, we review the relatively new field of protein structural phylogeny, an area of study in which information about the distribution of protein structures among different organisms is used to reconstruct the evolutionary relationships between them. Taken together, the work that we review presents an increasingly coherent picture of how these unique polymers have evolved over the course of life on Earth.

Markov random fields reveal an N-terminal double beta-propeller motif as part of a bacterial hybrid two-component sensor system

The recent explosion in newly sequenced bacterial genomes is outpacing the capacity of researchers to try to assign functional annotation to all the new proteins. Hence, computational methods that can help predict structural motifs provide increasingly important clues in helping to determine how these proteins might function. We introduce a Markov Random Field approach tailored for recognizing proteins that fold into mainly beta-structural motifs, and apply it to build recognizers for the beta-propeller shapes. As an application, we identify a potential class of hybrid two-component sensor proteins, that we predict contain a double-propeller domain.

Contact prediction in protein modeling: scoring, folding and refinement of coarse-grained models

BACKGROUND

Several different methods for contact prediction succeeded within the Sixth Critical Assessment of Techniques for Protein Structure Prediction (CASP6). The most relevant were non-local contact predictions for targets from the most difficult categories: fold recognition-analogy and new fold. Such contacts could provide valuable structural information in case a template structure cannot be found in the PDB.

RESULTS

We described comprehensive tests of the effectiveness of contact data in various aspects of de novo modeling with CABS, an algorithm which was used successfully in CASP6 by the Kolinski-Bujnicki group. We used the predicted contacts in a simple scoring function for the post-simulation ranking of protein models and as a soft bias in the folding simulations and in the fold-refinement procedure. The latter approach turned out to be the most successful. The CABS force field used in the Replica Exchange Monte Carlo simulations cooperated with the true contacts and discriminated the false ones, which resulted in an improvement of the majority of Kolinski-Bujnicki's protein models. In the modeling we tested different sets of predicted contact data submitted to the CASP6 server. According to our results, the best performing were the contacts with the accuracy balanced with the coverage, obtained either from the best two predictors only or by a consensus from as many predictors as possible.

CONCLUSION

Our tests have shown that theoretically predicted contacts can be very beneficial for protein structure prediction. Depending on the protein modeling method, a contact data set applied should be prepared with differently balanced coverage and accuracy of predicted contacts. Namely, high coverage of contact data is important for the model ranking and high accuracy for the folding simulations.

Assessment of CASP7 structure predictions for template free targets

In CASP7, protein structure prediction targets that lacked substantial similarity to a protein in the PDB at the time of assessment were considered to be free modeling targets (FM). We assessed predictions for 14 FM targets as well as four other targets that were deemed to be on the borderline between FM targets and template based modeling targets (TBM/FM). GDT_TS was used as one measure of model quality. Model quality was also assessed by visual inspection. Visual inspection was performed by three independent assessors who were blinded to GDT_TS scores and other quantitative measures of model quality. The best models by visual inspection tended to rank among the top few percent by GDT_TS, but were typically not the highest scoring models. Thus, visual inspection remains an essential component of assessment for FM targets. Overall, group TS020 (Baker) performed best, but success on individual targets was widely distributed among many groups. Among these other groups, TS024 and TS025 (Zhang and Zhang server) performed notably well without exceptionally large computing resources. This should be considered encouraging for future CASPs. There was a sense of progress in template FM relative to CASP6, but we were unable to demonstrate this progress objectively.

Superfamily assignments for the yeast proteome through integration of structure prediction with the gene ontology

Saccharomyces cerevisiae is one of the best-studied model organisms, yet the three-dimensional structure and molecular function of many yeast proteins remain unknown. Yeast proteins were parsed into 14,934 domains, and those lacking sequence similarity to proteins of known structure were folded using the Rosetta de novo structure prediction method on the World Community Grid. This structural data was integrated with process, component, and function annotations from the Saccharomyces Genome Database to assign yeast protein domains to SCOP superfamilies using a simple Bayesian approach. We have predicted the structure of 3,338 putative domains and assigned SCOP superfamily annotations to 581 of them. We have also assigned structural annotations to 7,094 predicted domains based on fold recognition and homology modeling methods. The domain predictions and structural information are available in an online database at http://rd.plos.org/10.1371_journal.pbio.0050076_01.

Accurate prediction of protein secondary structure and solvent accessibility by consensus combiners of sequence and structure information

BACKGROUND

Structural properties of proteins such as secondary structure and solvent accessibility contribute to three-dimensional structure prediction, not only in the ab initio case but also when homology information to known structures is available. Structural properties are also routinely used in protein analysis even when homology is available, largely because homology modelling is lower throughput than, say, secondary structure prediction. Nonetheless, predictors of secondary structure and solvent accessibility are virtually always ab initio.

RESULTS

Here we develop high-throughput machine learning systems for the prediction of protein secondary structure and solvent accessibility that exploit homology to proteins of known structure, where available, in the form of simple structural frequency profiles extracted from sets of PDB templates. We compare these systems to their state-of-the-art ab initio counterparts, and with a number of baselines in which secondary structures and solvent accessibilities are extracted directly from the templates. We show that structural information from templates greatly improves secondary structure and solvent accessibility prediction quality, and that, on average, the systems significantly enrich the information contained in the templates. For sequence similarity exceeding 30%, secondary structure prediction quality is approximately 90%, close to its theoretical maximum, and 2-class solvent accessibility roughly 85%. Gains are robust with respect to template selection noise, and significant for marginal sequence similarity and for short alignments, supporting the claim that these improved predictions may prove beneficial beyond the case in which clear homology is available.

CONCLUSION

The predictive system are publicly available at the address http://distill.ucd.ie.

Contact patterns between helices and strands of sheet define protein folding patterns

Comparing and classifying protein folding patterns allows organizing the known structures and enumerating possible protein structural patterns including those not yet observed. We capture the essence of protein folding patterns in a concise tableau representation based on the order and contact patterns of secondary structures: helices and strands of sheet. The tableaux are intelligible to both humans and computers. They provide a database, derived from the Protein Data Bank, mineable in studies of protein architecture. Using this database, we have: (i) determined statistical properties of secondary structure contacts in an unbiased set of protein domains from ASTRAL, (ii) observed that in 98% of cases, the tableau is a faithful representation of the folding pattern as classified in SCOP, (iii) demonstrated that to a large extent the local structure of proteins indicates their complete folding topology, and (iv) studied the use of the representation for fold identification.

Improved residue contact prediction using support vector machines and a large feature set

Background

Predicting protein residue-residue contacts is an important 2D prediction task. It is useful for ab initio structure prediction and understanding protein folding. In spite of steady progress over the past decade, contact prediction remains still largely unsolved.

Results

Here we develop a new contact map predictor (SVMcon) that uses support vector machines to predict medium- and long-range contacts. SVMcon integrates profiles, secondary structure, relative solvent accessibility, contact potentials, and other useful features. On the same test data set, SVMcon's accuracy is 4% higher than the latest version of the CMAPpro contact map predictor. SVMcon recently participated in the seventh edition of the Critical Assessment of Techniques for Protein Structure Prediction (CASP7) experiment and was evaluated along with seven other contact map predictors. SVMcon was ranked as one of the top predictors, yielding the second best coverage and accuracy for contacts with sequence separation >= 12 on 13 de novo domains.

Conclusion

We describe SVMcon, a new contact map predictor that uses SVMs and a large set of informative features. SVMcon yields good performance on medium- to long-range contact predictions and can be modularly incorporated into a structure prediction pipeline.

Fold recognition and accurate sequence-structure alignment of sequences directing beta-sheet proteins

The ability to predict structure from sequence is particularly important for toxins, virulence factors, allergens, cytokines, and other proteins of public health importance. Many such functions are represented in the parallel beta-helix and beta-trefoil families. A method using pairwise beta-strand interaction probabilities coupled with evolutionary information represented by sequence profiles is developed to tackle these problems for the beta-helix and beta-trefoil folds. The algorithm BetaWrapPro employs a "wrapping" component that may capture folding processes with an initiation stage followed by processive interaction of the sequence with the already-formed motifs. BetaWrapPro outperforms all previous motif recognition programs for these folds, recognizing the beta-helix with 100% sensitivity and 99.7% specificity and the beta-trefoil with 100% sensitivity and 92.5% specificity, in crossvalidation on a database of all nonredundant known positive and negative examples of these fold classes in the PDB. It additionally aligns 88% of residues for the beta-helices and 86% for the beta-trefoils accurately (within four residues of the exact position) to the structural template, which is then used with the side-chain packing program SCWRL to produce 3D structure predictions. One striking result has been the prediction of an unexpected parallel beta-helix structure for a pollen allergen, and its recent confirmation through solution of its structure. A Web server running BetaWrapPro is available and outputs putative PDB-style coordinates for sequences predicted to form the target folds.

Modular DAG-RNN architectures for assembling coarse protein structures

We develop and test machine learning methods for the prediction of coarse 3D protein structures, where a protein is represented by a set of rigid rods associated with its secondary structure elements (alpha-helices and beta-strands). First, we employ cascades of recursive neural networks derived from graphical models to predict the relative placements of segments. These are represented as discretized distance and angle maps, and the discretization levels are statistically inferred from a large and curated dataset. Coarse 3D folds of proteins are then assembled starting from topological information predicted in the first stage. Reconstruction is carried out by minimizing a cost function taking the form of a purely geometrical potential. We show that the proposed architecture outperforms simpler alternatives and can accurately predict binary and multiclass coarse maps. The reconstruction procedure proves to be fast and often leads to topologically correct coarse structures that could be exploited as a starting point for various protein modeling strategies. The fully integrated rod-shaped protein builder (predictor of contact maps + reconstruction algorithm) can be accessed at http://distill.ucd.ie/.

CASP6 assessment of contact prediction

Here we present the evaluation results of the Critical Assessment of Protein Structure Prediction (CASP6) contact prediction category. Contact prediction was assessed with standard measures well known in the field and the performance of specialist groups was evaluated alongside groups that submitted models with 3D coordinates. The evaluation was mainly focused on long range contact predictions for the set of new fold targets, although we analyzed predictions for all targets. Three groups with similar levels of accuracy and coverage performed a little better than the others. Comparisons of the predictions of the three best methods with those of CASP5/CAFASP3 suggested some improvement, although there were not enough targets in the comparisons to make this statistically significant.

A consensus procedure improving solvent accessibility prediction

Prediction methods of structural features in 1D represent a useful tool for the understanding of folding, classification, and function of proteins, and, in particular, for 3D structure prediction. Among the structural aspects characterizing a protein, solvent accessibility has received great attention in recent years. The available methods proposed for predicting accessibility have never considered the combination of the results deriving from different methods to construct a consensus prediction able to provide more reliable results. A consensus approach that increases prediction accuracy using three high-performance methods is described. The results of our method for three different protein data sets show that up to 3.0% improvement in prediction accuracy of solvent accessibility may be obtained by a consensus approach. The improvement also extends to the correlation coefficient. Application of our consensus approach to the accessibility prediction using only three prediction methods gives results better than single methods combined for consensus formation. Currently, the scarce availability of predictors with similar parameters defining solvent accessibility hinders the testing of other methods in our consensus procedure.

SimFold energy function for de novo protein structure prediction: consensus with Rosetta

Predicting protein tertiary structures by in silico folding is still very difficult for proteins that have new folds. Here, we developed a coarse-grained energy function, SimFold, for de novo structure prediction, performed a benchmark test of prediction with fragment assembly simulations for 38 test proteins, and proposed consensus prediction with Rosetta. The SimFold energy consists of many terms that take into account solvent-induced effects on the basis of physicochemical consideration. In the benchmark test, SimFold succeeded in predicting native structures within 6.5 A for 12 of 38 proteins; this success rate was the same as that by the publicly available version of Rosetta (ab initio version 1.2) run with default parameters. We investigated which energy terms in SimFold contribute to structure prediction performance, finding that the hydrophobic interaction is the most crucial for the prediction, whereas other sequence-specific terms have weak but positive roles. In the benchmark, well-predicted proteins by SimFold and by Rosetta were not the same for 5 of 12 proteins, which led us to introduce consensus prediction. With combined decoys, we succeeded in prediction for 16 proteins, four more than SimFold or Rosetta separately. For each of 38 proteins, structural ensembles generated by SimFold and by Rosetta were qualitatively compared by mapping sampled structural space onto two dimensions. For proteins of which one of the two methods succeeded and the other failed in prediction, the former had a less scattered ensemble located around the native. For proteins of which both methods succeeded in prediction, often two ensembles were mixed up.

Assessing a novel approach for predicting local 3D protein structures from sequence

We developed a novel approach for predicting local protein structure from sequence. It relies on the Hybrid Protein Model (HPM), an unsupervised clustering method we previously developed. This model learns three-dimensional protein fragments encoded into a structural alphabet of 16 protein blocks (PBs). Here, we focused on 11-residue fragments encoded as a series of seven PBs and used HPM to cluster them according to their local similarities. We thus built a library of 120 overlapping prototypes (mean fragments from each cluster), with good three-dimensional local approximation, i.e., a mean accuracy of 1.61 A Calpha root-mean-square distance. Our prediction method is intended to optimize the exploitation of the sequence-structure relations deduced from this library of long protein fragments. This was achieved by setting up a system of 120 experts, each defined by logistic regression to optimize the discrimination from sequence of a given prototype relative to the others. For a target sequence window, the experts computed probabilities of sequence-structure compatibility for the prototypes and ranked them, proposing the top scorers as structural candidates. Predictions were defined as successful when a prototype <2.5 A from the true local structure was found among those proposed. Our strategy yielded a prediction rate of 51.2% for an average of 4.2 candidates per sequence window. We also proposed a confidence index to estimate prediction quality. Our approach predicts from sequence alone and will thus provide valuable information for proteins without structural homologs. Candidates will also contribute to global structure prediction by fragment assembly.

Photochemical surface mapping of C14S-Sml1p for constrained computational modeling of protein structure

Photochemically generated hydroxyl radicals were used to map solvent-exposed regions in the C14S mutant of the protein Sml1p, a regulator of the ribonuclease reductase enzyme Rnr1p in Saccharomyces cerevisiae. By using high-performance mass spectrometry to characterize the oxidized peptides created by the hydroxyl radical reactions, amino acid solvent-accessibility data for native and denatured C14S Sml1p that revealed a solvent-excluding tertiary structure in the native state were obtained. The data on solvent accessibilities of various amino acids within the protein were then utilized to evaluate the de novo computational models generated by the HMMSTR/Rosetta server. The top five models initially generated by the server all disagreed with both published nuclear magnetic resonance (NMR) data and the solvent-accessibility data obtained in this study. A structural model adjusted to fit the previously reported NMR data satisfied most of the solvent-accessibility constraints. Through minor adjustment of the rotamers of two amino acid side chains for this latter structure, a model that not only provided a lower energy conformation but also completely satisfied previously reported data from NMR and tryptophan fluorescence measurements, in addition to the solvent-accessibility data presented here, was generated.

The emergence of scaling in sequence-based physical models of protein evolution

It has recently been discovered that many biological systems, when represented as graphs, exhibit a scale-free topology. One such system is the set of structural relationships among protein domains. The scale-free nature of this and other systems has previously been explained using network growth models that, although motivated by biological processes, do not explicitly consider the underlying physics or biology. In this work we explore a sequence-based model for the evolution protein structures and demonstrate that this model is able to recapitulate the scale-free nature observed in graphs of real protein structures. We find that this model also reproduces other statistical feature of the protein domain graph. This represents, to our knowledge, the first such microscopic, physics-based evolutionary model for a scale-free network of biological importance and as such has strong implications for our understanding of the evolution of protein structures and of other biological networks.

Prediction of functional tertiary interactions and intermolecular interfaces from primary sequence data

Given the availability of sequence information for many species, one can examine how the sequence of a gene varies among different organisms. This is accomplished by aligning the sequences and observing patterns of conservation, mutation and counter-mutation at different positions in the gene. Imbedded in these patterns is information on energetic coupling and macromolecular interactions, which can be deciphered by application of statistical algorithms. Here we report a robust approach for predicting interactions within (or between) any type of biopolymer, including proteins, RNAs and RNA-protein complexes. Rather than maximize the number of predictions, this approach is designed to detect a limited number of highly significant interactions, thereby providing accurate results from alignments that contain a modest number of sequences (20-60). The versatility and accuracy of the algorithm is demonstrated by the successful prediction of important intramolecular interactions within RNAs, modified RNAs, and proteins, as well as the prediction of RNA-protein and protein-protein interactions.

Protein structure prediction based on fragment assembly and parameter optimization

We propose a novel method for ab-initio prediction of protein tertiary structures based on the fragment assembly and global optimization. Fifteen residue long fragment libraries are constructed using the secondary structure prediction method PREDICT, and fragments in these libraries are assembled to generate full-length chains of a query protein. Tertiary structures of 50 to 100 conformations are obtained by minimizing an energy function for proteins, using the conformational space annealing method that enables one to sample diverse low-lying local minima of the energy. Then in order to enhance the performance of the prediction method, we optimize the linear parameters of the energy function, so that the native-like conformations become energetically more favorable than the non-native ones for proteins with known structures. We test the feasibility of the parameter optimization procedure by applying it to the training set consisting of three proteins: the 10-55 residue fragment of staphylococcal protein A (PDB ID 1bdd), a designed protein betanova, and 1fsd.

Prediction of protein tertiary structure using PROFESY, a novel method based on fragment assembly and conformational space annealing

A novel method for ab initio prediction of protein tertiary structures, PROFESY (PROFile Enumerating SYstem), is proposed. This method utilizes the secondary structure prediction information of a query sequence and the fragment assembly procedure based on global optimization. Fifteen-residue-long fragment libraries are constructed using the secondary structure prediction method PREDICT, and fragments in these libraries are assembled to generate full-length chains of a query protein. Tertiary structures of 50 to 100 conformations are obtained by minimizing an energy function for proteins, using the conformational space annealing method that enables one to sample diverse low-lying local minima of the energy. We apply PROFESY for benchmark tests to proteins with known structures to demonstrate its feasibility. In addition, we participated in CASP5 and applied PROFESY to four new-fold targets for blind prediction. The results are quite promising, despite the fact that PROFESY was in its early stages of development. In particular, PROFESY successfully provided us the best model-one structure for the target T0161.

Evaluation of the suitability of free-energy minimization using nearest-neighbor energy parameters for RNA secondary structure prediction

BACKGROUND

A detailed understanding of an RNA's correct secondary and tertiary structure is crucial to understanding its function and mechanism in the cell. Free energy minimization with energy parameters based on the nearest-neighbor model and comparative analysis are the primary methods for predicting an RNA's secondary structure from its sequence. Version 3.1 of Mfold has been available since 1999. This version contains an expanded sequence dependence of energy parameters and the ability to incorporate coaxial stacking into free energy calculations. We test Mfold 3.1 by performing the largest and most phylogenetically diverse comparison of rRNA and tRNA structures predicted by comparative analysis and Mfold, and we use the results of our tests on 16S and 23S rRNA sequences to assess the improvement between Mfold 2.3 and Mfold 3.1.

RESULTS

The average prediction accuracy for a 16S or 23S rRNA sequence with Mfold 3.1 is 41%, while the prediction accuracies for the majority of 16S and 23S rRNA structures tested are between 20% and 60%, with some having less than 20% prediction accuracy. The average prediction accuracy was 71% for 5S rRNA and 69% for tRNA. The majority of the 5S rRNA and tRNA sequences have prediction accuracies greater than 60%. The prediction accuracy of 16S rRNA base-pairs decreases exponentially as the number of nucleotides intervening between the 5' and 3' halves of the base-pair increases.

CONCLUSION

Our analysis indicates that the current set of nearest-neighbor energy parameters in conjunction with the Mfold folding algorithm are unable to consistently and reliably predict an RNA's correct secondary structure. For 16S or 23S rRNA structure prediction, Mfold 3.1 offers little improvement over Mfold 2.3. However, the nearest-neighbor energy parameters do work well for shorter RNA sequences such as tRNA or 5S rRNA, or for larger rRNAs when the contact distance between the base-pairs is less than 100 nucleotides.

Efficient methods for filtering and ranking fragments for the prediction of structurally variable regions in proteins

The prediction of protein 3D structures close to insertions and deletions or, more generally, loop prediction, is still one of the major challenges in homology modeling projects. In this article, we developed ranking criteria and selection filters to improve knowledge-based loop predictions. These criteria were developed and optimized for a test data set containing 678 insertions and deletions. The examples are, in principle, predictable from the used loop database with an RMSD < 1 A and represent realistic modeling situations. Four noncorrelated criteria for the selection of fragments are evaluated. A fast prefilter compares the distance between the anchor groups in the template protein with the stems of the fragments. The RMSD of the anchor groups is used for fitting and ranking of the selected loop candidates. After fitting, repulsive close contacts of loop candidates with the template protein are used for filtering, and fragments with backbone torsion angles, which are unfavorable according to a knowledge-based potential, are eliminated. By the combined application of these filter criteria to the test set, it was possible to increase the percentage of predictions with a global RMSD < 1 A to over 50% among the first five ranks, with average global RMSD values for the first rank candidate that are between 1.3 and 2.2 A for different loop lengths. Compared to other examples described in the literature, our large numbers of test cases are not self-predictions, where loops are placed in a protein after a peptide loop has been cut out, but are attempts to predict structural changes that occur in evolution when a protein is affected by insertions and deletions.

Critical assessment of methods of protein structure prediction (CASP)-round V

This article provides an introduction to the special issue of the journal Proteins dedicated to the fifth CASP experiment to assess the state of the art in protein structure prediction. The article describes the conduct, the categories of prediction, and the evaluation and assessment procedures of the experiment. A brief summary of progress over the five CASP experiments is provided. Related developments in the field are also described.

Protein evolution within a structural space

Understanding of the evolutionary origins of protein structures represents a key component of the understanding of molecular evolution as a whole. Here we seek to elucidate how the features of an underlying protein structural "space" might impact protein structural evolution. We approach this question using lattice polymers as a completely characterized model of this space. We develop a measure of structural comparison of lattice structures that is analogous to the one used to understand structural similarities between real proteins. We use this measure of structural relatedness to create a graph of lattice structures and compare this graph (in which nodes are lattice structures and edges are defined using structural similarity) to the graph obtained for real protein structures. We find that the graph obtained from all compact lattice structures exhibits a distribution of structural neighbors per node consistent with a random graph. We also find that subgraphs of 3500 nodes chosen either at random or according to physical constraints also represent random graphs. We develop a divergent evolution model based on the lattice space which produces graphs that, within certain parameter regimes, recapitulate the scale-free behavior observed in similar graphs of real protein structures.

Predictions without templates: New folds, secondary structure, and contacts in CASP5

We present the assessment of CASP5 predictions in the new fold category. For coordinate predictions, we considered five targets with new folds and eight lying on the fold recognition borderline. We performed detailed visual and numerical comparisons between predicted and experimental structures to assess prediction accuracy. The two procedures largely agreed, but the visual inspection identified instances where metrics, such as GDT_TS, ranked what we considered incorrect predictions highly. We found the quality of the best predictions to be very good: for nearly every target at least one group predicted a structure close to the correct one. However, selection of the best of five models is still problematic. The group of David Baker once again proved to be best overall, with many individual highlights. However, high quality and consistency were also seen from others, suggesting that the community is moving toward general procedures to predict accurate structures for proteins showing no resemblance to anything seen before. Predictions for secondary structure showed at best limited progress since CASP4. The number of targets is probably too small to spot differences in performance between methods, suggesting that such predictions might be better evaluated with schemes involving more proteins. For contact predictions, accuracies are still low, although there were several instances of accurate and useful contacts predicted de novo, and new approaches hint at future progress.

Deciphering a novel thioredoxin-like fold family

Sequence--and structure-based searching strategies have proven useful in the identification of remote homologs and have facilitated both structural and functional predictions of many uncharacterized protein families. We implement these strategies to predict the structure of and to classify a previously uncharacterized cluster of orthologs (COG3019) in the thioredoxin-like fold superfamily. The results of each searching method indicate that thioltransferases are the closest structural family to COG3019. We substantiate this conclusion using the ab initio structure prediction method rosetta, which generates a thioredoxin-like fold similar to that of the glutaredoxin-like thioltransferase (NrdH) for a COG3019 target sequence. This structural model contains the thiol-redox functional motif CYS-X-X-CYS in close proximity to other absolutely conserved COG3019 residues, defining a novel thioredoxin-like active site that potentially binds metal ions. Finally, the rosetta-derived model structure assists us in assembling a global multiple-sequence alignment of COG3019 with two other thioredoxin-like fold families, the thioltransferases and the bacterial arsenate reductases (ArsC).

Hybrid global optimization algorithms for protein structure prediction: alternating hybrids

Hybrid global optimization methods attempt to combine the beneficial features of two or more algorithms, and can be powerful methods for solving challenging nonconvex optimization problems. In this paper, novel classes of hybrid global optimization methods, termed alternating hybrids, are introduced for application as a tool in treating the peptide and protein structure prediction problems. In particular, these new optimization methods take the form of hybrids between a deterministic global optimization algorithm, the alphaBB, and a stochastically based method, conformational space annealing (CSA). The alphaBB method, as a theoretically proven global optimization approach, exhibits consistency, as it guarantees convergence to the global minimum for twice-continuously differentiable constrained nonlinear programming problems, but can benefit from computationally related enhancements. On the other hand, the independent CSA algorithm is highly efficient, though the method lacks theoretical guarantees of convergence. Furthermore, both the alphaBB method and the CSA method are found to identify ensembles of low-energy conformers, an important feature for determining the true free energy minimum of the system. The proposed hybrid methods combine the desirable features of efficiency and consistency, thus enabling the accurate prediction of the structures of larger peptides. Computational studies for met-enkephalin and melittin, employing sequential and parallel computing frameworks, demonstrate the promise for these proposed hybrid methods.

De novo prediction of three-dimensional structures for major protein families

We use the Rosetta de novo structure prediction method to produce three-dimensional structure models for all Pfam-A sequence families with average length under 150 residues and no link to any protein of known structure. To estimate the reliability of the predictions, the method was calibrated on 131 proteins of known structure. For approximately 60% of the proteins one of the top five models was correctly predicted for 50 or more residues, and for approximately 35%, the correct SCOP superfamily was identified in a structure-based search of the Protein Data Bank using one of the models. This performance is consistent with results from the fourth critical assessment of structure prediction (CASP4). Correct and incorrect predictions could be partially distinguished using a confidence function based on a combination of simulation convergence, protein length and the similarity of a given structure prediction to known protein structures. While the limited accuracy and reliability of the method precludes definitive conclusions, the Pfam models provide the only tertiary structure information available for the 12% of publicly available sequences represented by these large protein families.