Assessment of protein side‐chain conformation prediction methods in different residue environments

Computational prediction of side‐chain conformation is an important component of protein structure prediction. Accurate side‐chain prediction is crucial for practical applications of protein structure models that need atomic‐detailed resolution such as protein and ligand design. We evaluated the accuracy of eight side‐chain prediction methods in reproducing the side‐chain conformations of experimentally solved structures deposited to the Protein Data Bank. Prediction accuracy was evaluated for a total of four different structural environments (buried, surface, interface, and membrane‐spanning) in three different protein types (monomeric, multimeric, and membrane). Overall, the highest accuracy was observed for buried residues in monomeric and multimeric proteins. Notably, side‐chains at protein interfaces and membrane‐spanning regions were better predicted than surface residues even though the methods did not all use multimeric and membrane proteins for training. Thus, we conclude that the current methods are as practically useful for modeling protein docking interfaces and membrane‐spanning regions as for modeling monomers. Proteins 2014; 82:1971–1984. © 2014 Wiley Periodicals, Inc.     

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.     

Support vector machines for the classification and prediction of beta-turn types.

The support vector machines (SVMs) method is proposed because it can reflect the sequence-coupling effect for a tetrapeptide in not only a beta-turn or non-beta-turn, but also in different types of beta-turn. The results of the model for 6022 tetrapeptides indicate that the rates of self-consistency for beta-turn types I, I', II, II', VI and VIII and non-beta-turns are 99.92%, 96.8%, 98.02%, 97.75%, 100%, 97.19% and 100%, respectively. Using these training data, the rate of correct prediction by the SVMs for a given protein: rubredoxin (54 residues. 51 tetrapeptides) which includes 12 beta-turn type I tetrapeptides, 1 beta-turn type II tetrapeptide and 38 non-beta-turns reached 82.4%. The high quality of prediction of the SVMs implies that the formation of different beta-turn types or non-beta-turns is considerably correlated with the sequence of a tetrapeptide. The SVMs can save CPU time and avoid the overfitting problem compared with the neural network method.     

Optimal data collection for correlated mutation analysis

The main objective of correlated mutation analysis (CMA) is to predict intraprotein residue-residue interactions from sequence alone. Despite considerable progress in algorithms and computer capabilities, the performance of CMA methods remains quite low. Here we examine whether, and to what extent, the quality of CMA methods depends on the sequences that are included in the multiple sequence alignment (MSA). The results revealed a strong correlation between the number of homologs in an MSA and CMA prediction strength. Furthermore, many of the current methods include only orthologs in the MSA, we found that it is beneficial to include both orthologs and paralogs in the MSA. Remarkably, even remote homologs contribute to the improved accuracy. Based on our findings we put forward an automated data collection procedure, with a minimal coverage of 50% between the query protein and its orthologs and paralogs. This procedure improves accuracy even in the absence of manual curation. In this era of massive sequencing and exploding sequence data, our results suggest that correlated mutation-based methods have not reached their inherent performance limitations and that the role of CMA in structural biology is far from being fulfilled.     

Exploring alternate states and oligomerization preferences of coiled‐coils by de novo structure modeling

Homomeric coiled‐coils can self‐assemble into a wide range of structural states with different helix topologies and oligomeric states. In this study, we have combined de novo structure modeling with stability calculations to simultaneously predict structure and oligomeric states of homomeric coiled‐coils. For dimers an asymmetric modeling protocol was developed. Modeling without symmetry constraints showed that backbone asymmetry is important for the formation of parallel dimeric coiled‐coils. Collectively, our results demonstrate that high‐resolution structure of coiled‐coils, as well as parallel and antiparallel orientations of dimers and tetramers, can be accurately predicted from sequence. De novo modeling was also used to generate models of competing oligomeric states, which were used to compare stabilities and thus predict the native stoichiometry from sequence. In a benchmark set of 33 coiled‐coil sequences, forming dimers to pentamers, up to 70% of the oligomeric states could be correctly predicted. The calculations demonstrated that the free energy of helix folding could be an important factor for determining stability and oligomeric state of homomeric coiled‐coils. The computational methods developed here should be broadly applicable to studies of sequence‐structure relationships in coiled‐coils and the design of higher order assemblies with improved oligomerization specificity. Proteins 2015; 83:235–247. © 2014 Wiley Periodicals, Inc.     

Analysis of a data set of paired uncomplexed protein structures: new metrics for side-chain flexibility and model evaluation

We compiled and analyzed a data set of paired protein structures containing proteins for which multiple high-quality uncomplexed atomic structures were available in the Protein Data Bank. Side-chain flexibility was quantified, yielding a set of residue- and environment-specific confidence levels describing the range of motion around chi1 and chi2 angles. As expected, buried residues were inflexible, adopting similar conformations in different crystal structure analyses. Ile, Thr, Asn, Asp, and the large aromatics also showed limited flexibility when exposed on the protein surface, whereas exposed Ser, Lys, Arg, Met, Gln, and Glu residues were very flexible. This information is different from and complementary to the information available from rotamer surveys. The confidence levels are useful for assessing the significance of observed side-chain motion and estimating the extent of side-chain motion in protein structure prediction. We compare the performance of a simple 40 degrees threshold with these quantitative confidence levels in a critical evaluation of side-chain prediction with the program SCWRL.     

Prediction errors for state occupation and transition probabilities in multi‐state models

In this paper, we consider the estimation of prediction errors for state occupation probabilities and transition probabilities for multistate time‐to‐event data. We study prediction errors based on the Brier score and on the Kullback–Leibler score and prove their properness. In the presence of right‐censored data, two classes of estimators, based on inverse probability weighting and pseudo‐values, respectively, are proposed, and consistency properties of the proposed estimators are investigated. The second part of the paper is devoted to the estimation of dynamic prediction errors for state occupation probabilities for multistate models, conditional on being alive, and for transition probabilities. Cross‐validated versions are proposed. Our methods are illustrated on the CSL1 randomized clinical trial comparing prednisone versus placebo for liver cirrhosis patients.     

High‐accuracy modeling of antibody structures by a search for minimum‐energy recombination of backbone fragments

Current methods for antibody structure prediction rely on sequence homology to known structures. Although this strategy often yields accurate predictions, models can be stereo‐chemically strained. Here, we present a fully automated algorithm, called AbPredict, that disregards sequence homology, and instead uses a Monte Carlo search for low‐energy conformations built from backbone segments and rigid‐body orientations that appear in antibody molecular structures. We find cases where AbPredict selects accurate loop templates with sequence identity as low as 10%, whereas the template of highest sequence identity diverges substantially from the query's conformation. Accordingly, in several cases reported in the recent Antibody Modeling Assessment benchmark, AbPredict models were more accurate than those from any participant, and the models' stereo‐chemical quality was consistently high. Furthermore, in two blind cases provided to us by crystallographers prior to structure determination, the method achieved <1.5 Ångstrom overall backbone accuracy. Accurate modeling of unstrained antibody structures will enable design and engineering of improved binders for biomedical research directly from sequence. Proteins 2016; 85:30–38. © 2016 Wiley Periodicals, Inc.     

Microbial genomes have over 72% structure assignment by the threading algorithm PROSPECTOR_Q

The genome scale threading of five complete microbial genomes is revisited using our state-of-the-art threading algorithm, PROSPECTOR_Q. Considering that structure assignment to an ORF could be useful for predicting biochemical function as well as for analyzing pathways, it is important to assess the current status of genome scale threading. The fraction of ORFs to which we could assign protein structures with a reasonably good confidence level to each genome sequences is over 72%, which is significantly higher than earlier studies. Using the assigned structures, we have predicted the function of several ORFs through "single-function" template structures, obtained from an analysis of the relationship between protein fold and function. The fold distribution of the genomes and the effect of the number of homologous sequences on structure assignment are also discussed.     

Genome‐scale prediction of proteins with long intrinsically disordered regions

Proteins with long disordered regions (LDRs), defined as having 30 or more consecutive disordered residues, are abundant in eukaryotes, and these regions are recognized as a distinct class of biologically functional domains. LDRs facilitate various cellular functions and are important for target selection in structural genomics. Motivated by the lack of methods that directly predict proteins with LDRs, we designed Super‐fast predictor of proteins with Long Intrinsically DisordERed regions (SLIDER). SLIDER utilizes logistic regression that takes an empirically chosen set of numerical features, which consider selected physicochemical properties of amino acids, sequence complexity, and amino acid composition, as its inputs. Empirical tests show that SLIDER offers competitive predictive performance combined with low computational cost. It outperforms, by at least a modest margin, a comprehensive set of modern disorder predictors (that can indirectly predict LDRs) and is 16 times faster compared to the best currently available disorder predictor. Utilizing our time‐efficient predictor, we characterized abundance and functional roles of proteins with LDRs over 110 eukaryotic proteomes. Similar to related studies, we found that eukaryotes have many (on average 30.3%) proteins with LDRs with majority of proteomes having between 25 and 40%, where higher abundance is characteristic to proteomes that have larger proteins. Our first‐of‐its‐kind large‐scale functional analysis shows that these proteins are enriched in a number of cellular functions and processes including certain binding events, regulation of catalytic activities, cellular component organization, biogenesis, biological regulation, and some metabolic and developmental processes. A webserver that implements SLIDER is available at http://biomine.ece.ualberta.ca/SLIDER/ .Proteins 2014; 82:145–158. © 2013 Wiley Periodicals, Inc.     

Prediction of protein–protein interaction sites from weakly homologous template structures using meta‐threading and machine learning

The identification of protein–protein interactions is vital for understanding protein function, elucidating interaction mechanisms, and for practical applications in drug discovery. With the exponentially growing protein sequence data, fully automated computational methods that predict interactions between proteins are becoming essential components of system‐level function inference. A thorough analysis of protein complex structures demonstrated that binding site locations as well as the interfacial geometry are highly conserved across evolutionarily related proteins. Because the conformational space of protein–protein interactions is highly covered by experimental structures, sensitive protein threading techniques can be used to identify suitable templates for the accurate prediction of interfacial residues. Toward this goal, we developed eFindSite^PPI, an algorithm that uses the three‐dimensional structure of a target protein, evolutionarily remotely related templates and machine learning techniques to predict binding residues. Using crystal structures, the average sensitivity (specificity) of eFindSite^PPI in interfacial residue prediction is 0.46 (0.92). For weakly homologous protein models, these values only slightly decrease to 0.40–0.43 (0.91–0.92) demonstrating that eFindSite^PPI performs well not only using experimental data but also tolerates structural imperfections in computer‐generated structures. In addition, eFindSite^PPI detects specific molecular interactions at the interface; for instance, it correctly predicts approximately one half of hydrogen bonds and aromatic interactions, as well as one third of salt bridges and hydrophobic contacts. Comparative benchmarks against several dimer datasets show that eFindSite^PPI outperforms other methods for protein‐binding residue prediction. It also features a carefully tuned confidence estimation system, which is particularly useful in large‐scale applications using raw genomic data. eFindSite^PPI is freely available to the academic community at http://www.brylinski.org/efindsiteppi . Copyright © 2014 John Wiley & Sons, Ltd.     

Prediction of ligand-binding sites of proteins by molecular docking calculation for a random ligand library

A new approach to predicting the ligand-binding sites of proteins was developed, using protein-ligand docking computation. In this method, many compounds in a random library are docked onto the whole protein surface. We assumed that the true ligand-binding site would exhibit stronger affinity to the compounds in the random library than the other sites, even if the random library did not include the ligand corresponding to the true binding site. We also assumed that the affinity of the true ligand-binding site would be correlated to the docking scores of the compounds in the random library, if the ligand-binding site was correctly predicted. We call this method the molecular-docking binding-site finding (MolSite) method. The MolSite method was applied to 89 known protein-ligand complex structures extracted from the Protein Data Bank, and it predicted the correct binding sites with about 80-99% accuracy, when only the single top-ranked site was adopted. In addition, the average docking score was weakly correlated to the experimental protein-ligand binding free energy, with a correlation coefficient of 0.44.     

Improving threading algorithms for remote homology modeling by combining fragment and template comparisons

In this work, we develop a method called fragment comparison and the template comparison (FTCOM) for assessing the global quality of protein structural models for targets of medium and hard difficulty (remote homology) produced by structure prediction approaches such as threading or ab initio structure prediction. FTCOM requires the C_α coordinates of full length models and assesses model quality based on fragment comparison and a score derived from comparison of the model to top threading templates. On a set of 361 medium/hard targets, FTCOM was applied to and assessed for its ability to improve on the results from the SP³, SPARKS, PROSPECTOR_3, and PRO‐SP³‐TASSER threading algorithms. The average TM‐score improves by 5–10% for the first selected model by the new method over models obtained by the original selection procedure in the respective threading methods. Moreover, the number of foldable targets (TM‐score ≥ 0.4) increases from least 7.6% for SP³ to 54% for SPARKS. Thus, FTCOM is a promising approach to template selection. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Loop prediction for a GPCR homology model: Algorithms and results

We present loop structure prediction results of the intracellular and extracellular loops of four G‐protein‐coupled receptors (GPCRs): bovine rhodopsin (bRh), the turkey β1‐adrenergic (β1Ar), the human β2‐adrenergic (β2Ar) and the human A2a adenosine receptor (A2Ar) in perturbed environments. We used the protein local optimization program, which builds thousands of loop candidates by sampling rotamer states of the loops' constituent amino acids. The candidate loops are discriminated between with our physics‐based, all‐atom energy function, which is based on the OPLS force field with implicit solvent and several correction terms. For relevant cases, explicit membrane molecules are included to simulate the effect of the membrane on loop structure. We also discuss a new sampling algorithm that divides phase space into different regions, allowing more thorough sampling of long loops that greatly improves results. In the first half of the paper, loop prediction is done with the GPCRs' transmembrane domains fixed in their crystallographic positions, while the loops are built one‐by‐one. Side chains near the loops are also in non‐native conformations. The second half describes a full homology model of β2Ar using β1Ar as a template. No information about the crystal structure of β2Ar was used to build this homology model. We are able to capture the architecture of short loops and the very long second extracellular loop, which is key for ligand binding. We believe this the first successful example of an RMSD validated, physics‐based loop prediction in the context of a GPCR homology model. Proteins 2013. © 2012 Wiley Periodicals, Inc.     

Statistical mechanics-based method to extract atomic distance-dependent potentials from protein structures

In this study, we have developed a statistical mechanics-based iterative method to extract statistical atomic interaction potentials from known, nonredundant protein structures. Our method circumvents the long-standing reference state problem in deriving traditional knowledge-based scoring functions, by using rapid iterations through a physical, global convergence function. The rapid convergence of this physics-based method, unlike other parameter optimization methods, warrants the feasibility of deriving distance-dependent, all-atom statistical potentials to keep the scoring accuracy. The derived potentials, referred to as ITScore/Pro, have been validated using three diverse benchmarks: the high-resolution decoy set, the AMBER benchmark decoy set, and the CASP8 decoy set. Significant improvement in performance has been achieved. Finally, comparisons between the potentials of our model and potentials of a knowledge-based scoring function with a randomized reference state have revealed the reason for the better performance of our scoring function, which could provide useful insight into the development of other physical scoring functions. The potentials developed in this study are generally applicable for structural selection in protein structure prediction.     

A novel method for predicting and using distance constraints of high accuracy for refining protein structure prediction

The principal bottleneck in protein structure prediction is the refinement of models from lower accuracies to the resolution observed by experiment. We developed a novel constraints‐based refinement method that identifies a high number of accurate input constraints from initial models and rebuilds them using restrained torsion angle dynamics (rTAD). We previously created a Bayesian statistics‐based residue‐specific all‐atom probability discriminatory function (RAPDF) to discriminate native‐like models by measuring the probability of accuracy for atom type distances within a given model. Here, we exploit RAPDF to score (i.e., filter) constraints from initial predictions that may or may not be close to a native‐like state, obtain consensus of top scoring constraints amongst five initial models, and compile sets with no redundant residue pair constraints. We find that this method consistently produces a large and highly accurate set of distance constraints from which to build refinement models. We further optimize the balance between accuracy and coverage of constraints by producing multiple structure sets using different constraint distance cutoffs, and note that the cutoff governs spatially near versus distant effects in model generation. This complete procedure of deriving distance constraints for rTAD simulations improves the quality of initial predictions significantly in all cases evaluated by us. Our procedure represents a significant step in solving the protein structure prediction and refinement problem, by enabling the use of consensus constraints, RAPDF, and rTAD for protein structure modeling and refinement. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

COG4849 is a novel family of nucleotidyltransferases

Only a minority of currently known protein families is characterized structurally. This makes homology-based structure modeling an essential instrument that can be viewed as the first approximation to experimental determination of protein structure. Using sequence similarity searches, we detected a distant similarity between a family of uncharacterized hypothetical proteins, COG4849, and the family of tRNA nucleotidyltransferases. The suggested remote homology between the N-terminal domain of COG4849 and the catalytic domain of tRNA nucleotidyltransferase was further supported by comparison of sequence profiles, methods for fold recognition and structure modeling. The combined multiple alignment of the two families reveals shared conservation of functionally important motifs and suggests the similarity in catalytic mechanisms of the performed reactions. Our results suggest that (i) the N-terminal domain of proteins from COG4849 shares structural similarity with the catalytic domain of tRNA nucleotidyltransferase, and (ii) this domain catalyzes the nucleotidyl transfer reaction involving two metal ions.     

Physicochemical explanation of peptide binding to HLA-A*0201 major histocompatibility complex: a three-dimensional quantitative structure-activity relationship study

A three-dimensional quantitative structure-activity relationship method for the prediction of peptide binding affinities to the MHC class I molecule HLA-A*0201 was developed by applying the CoMSIA technique on a set of 266 peptides. To increase the self consistency of the initial CoMSIA model, the poorly predicted peptides were excluded from the training set in a stepwise manner and then included in the study as a test set. The final model, based on 236 peptides and considering the steric, electrostatic, hydrophobic, hydrogen bond donor, and hydrogen bond acceptor fields, had q2 = 0.683 and r2 = 0.891. The stability of this model was proven by cross-validations in two and five groups and by a bootstrap analysis of the non-cross-validated model. The residuals between the experimental pIC50 (-logIC50) values and those calculated by "leave-one-out" cross-validation were analyzed. According to the best model, 63.2% of the peptides were predicted with /residuals/ < or = 0.5 log unit; 29.3% with 1.0 < or = /residuals/ < 0.5; and 7.5% with /residuals/ > 1.0 log unit. The mean /residual/ value was 0.489. The coefficient contour maps identify the physicochemical property requirements at each position in the peptide molecule and suggest amino acid sequences for high-affinity binding to the HLA-A*0201 molecule.