Protein-protein docking predictions for the CAPRI experiment

We predicted structures for all seven targets in the CAPRI experiment using a new method in development at the time of the challenge. The technique includes a low-resolution rigid body Monte Carlo search followed by high-resolution refinement with side-chain conformational changes and rigid body minimization. Decoys (approximately 10(6) per target) were discriminated using a scoring function including van der Waals and solvation interactions, hydrogen bonding, residue-residue pair statistics, and rotamer probabilities. Decoys were ranked, clustered, manually inspected, and selected. The top ranked model for target 6 predicted the experimental structure to 1.5 A RMSD and included 48 of 65 correct residue-residue contacts. Target 7 was predicted at 5.3 A RMSD with 22 of 37 correct residue-residue contacts using a homology model from a known complex structure. Using a preliminary version of the protocol in round 1, target 1 was predicted within 8.8 A although few contacts were correct. For targets 2 and 3, the interface locations and a small fraction of the contacts were correctly identified.     

Evaluation of the 3D-Dock protein docking suite in rounds 1 and 2 of the CAPRI blind trial

The 3D-Dock suite of programs has been used to make predictions for the seven targets in rounds 1 and 2 of the CAPRI method evaluation exercise. Some correct contacts were obtained in at least one prediction for four of seven targets. Target 06 was predicted very well, with an RMSD of the ligand after superimposition of the receptor of only 0.77 A. We investigate the performance of the various stages of the method, with the aim of finding where improvements need to be made, and in particular whether the manual interventions that were made were essential, and whether results of the level of accuracy obtained for target 06 may be expected with confidence.     

Protein-protein docking in CAPRI using ATTRACT to account for global and local flexibility

A reduced protein model combined with a systematic docking approach has been employed to predict protein-protein complex structures in CAPRI rounds 6-11. The docking approach termed ATTRACT is based on energy minimization in translational and rotational degrees of freedom of one protein with respect to the second protein starting from many thousand initial protein partner placements. It also allows for approximate inclusion of global flexibility of protein partners during systematic docking by conformational relaxation of the partner proteins in precalculated soft collective backbone degrees of freedom. We have submitted models for six targets, achieved acceptable docking solutions for two targets, and predicted >20% correct contacts for five targets. Possible improvements of the docking approach in particular at the scoring and refinement steps are discussed.     

Docking prediction using biological information, ZDOCK sampling technique, and clustering guided by the DFIRE statistical energy function

We entered the CAPRI experiment during the middle of Round 4 and have submitted predictions for all 6 targets released since then. We used the following procedures for docking prediction: (1) the identification of possible binding region(s) of a target based on known biological information, (2) rigid-body sampling around the binding region(s) by using the docking program ZDOCK, (3) ranking of the sampled complex conformations by employing the DFIRE-based statistical energy function, (4) clustering based on pairwise root-mean-square distance and the DFIRE energy, and (5) manual inspection and relaxation of the side-chain conformations of the top-ranked structures by geometric constraint. Reasonable predictions were made for 4 of the 6 targets. The best fraction of native contacts within the top 10 models are 89.1% for Target 12, 54.3% for Target 13, 29.3% for Target 14, and 94.1% for Target 18. The origin of successes and failures is discussed. .     

Performance of human and server prediction in CAPRI rounds 38-45

CAPRI challenges offer a variety of blind tests for protein-protein interaction prediction. In CAPRI Rounds 38-45, we generated a set of putative binding modes for each target with an FFT-based docking algorithm, and then scored and ranked these binding modes with a proprietary scoring function, ITScorePP. We have also developed a novel web server, Rebipp. The algorithm utilizes information retrieval to identify relevant biological information to significantly reduce the search space for a particular protein. In parallel, we have also constructed a GPU-based docking server, MDockPP, for protein-protein complex structure prediction. Here, the performance of our protocol in CAPRI rounds 38-45 is reported, which include 16 docking and scoring targets. Among them, three targets contain multiple interfaces: Targets 124, 125, and 136 have 2, 4, and 3 interfaces, respectively. In the predictor experiments, we predicted correct binding modes for nine targets, including one high-accuracy interface, six medium-accuracy binding modes, and six acceptable-accuracy binding modes. For the docking server prediction experiments, we predicted correct binding modes for eight targets, including one high-accuracy, three medium-accuracy, and five acceptable-accuracy binding modes.     

Prediction of the unknown: inspiring experience with the CAPRI experiment

We submitted predictions for all seven targets in the CAPRI experiment. For four targets, our submitted models included acceptable, medium accuracy predictions of the structures of the complexes, and for a fifth target we identified the location of the binding site of one of the molecules. We used a weighted-geometric docking algorithm in which contacts involving specified parts of the surfaces of either one or both molecules were up-weighted or down-weighted. The weights were based on available structural and biochemical data or on sequence analyses. The weighted-geometric docking proved very useful for five targets, improving the complementarity scores and the ranks of the nearly correct solutions, as well as their statistical significance. In addition, the weighted-geometric docking promoted formation of clusters of similar solutions, which include more accurate predictions.     

Data-driven docking: HADDOCK's adventures in CAPRI

We have shown previously that given high-resolution structures of the unbound molecules, structure determination of protein complexes is possible by including biochemical and/or biophysical data as highly ambiguous distance restraints in a docking approach. We applied this method, implemented in the HADDOCK (High Ambiguity Driven DOCKing) package (Dominguez et al., J Am Chem Soc 2003;125:1731-1737), to the targets in the fourth and fifth rounds of CAPRI. Here we describe our results and analyze them in detail. Special attention is given to the role of flexibility in our docking method and the way in which this improves the docking results. We describe extensions to our approach that were developed as a direct result of our participation in CAPRI. In addition to experimental information, we also included interface residue predictions from PPISP (Protein-Protein Interaction Site Predictor; Zhou and Shan, Proteins 2001;44:336-343), a neural network method. Using HADDOCK we were able to generate acceptable structures for 6 of the 8 targets, and to submit at least 1 acceptable structure for 5 of them. Of these 5 submissions, 3 were of medium quality (Targets 10, 11, and 15) and 2 of high quality (Targets 13 and 14). In all cases, predictions were obtained containing at least 40% of the correct epitope at the interface for both ligand and receptor simultaneously.     

A comprehensive assessment of sequence-based and template-based methods for protein contact prediction

Motivation: Pair-wise residue-residue contacts in proteins canbe predicted from both threading templates and sequence-basedmachine learning. However, most structure modeling approachesonly use the template-based contact predictions in guiding thesimulations; this is partly because the sequence-based contactpredictions are usually considered to be less accurate thanthat by threading. With the rapid progress in sequence databasesand machine-learning techniques, it is necessary to have a detailedand comprehensive assessment of the contact-prediction methodsin different template conditions. Results: We develop two methods for protein-contact predictions:SVM-SEQ is a sequence-based machine learning approach whichtrains a variety of sequence-derived features on contact maps;SVM-LOMETS collects consensus contact predictions from multiplethreading templates. We test both methods on the same set of554 proteins which are categorized into ‘Easy’,‘Medium’, ‘Hard’ and ‘Very Hard’targets based on the evolutionary and structural distance betweentemplates and targets. For the Easy and Medium targets, SVM-LOMETSobviously outperforms SVM-SEQ; but for the Hard and Very Hardtargets, the accuracy of the SVM-SEQ predictions is higher thanthat of SVM-LOMETS by 12–25%. If we combine the SVM-SEQand SVM-LOMETS predictions together, the total number of correctlypredicted contacts in the Hard proteins will increase by morethan 60% (or 70% for the long-range contact with a sequenceseparation 24), compared with SVM-LOMETS alone. The advantageof SVM-SEQ is also shown in the CASP7 free modeling targetswhere the SVM-SEQ is around four times more accurate than SVM-LOMETSin the long-range contact prediction. These data demonstratethat the state-of-the-art sequence-based contact predictionhas reached a level which may be helpful in assisting tertiarystructure modeling for the targets which do not have close structuretemplates. The maximum yield should be obtained by the combinationof both sequence- and template-based predictions. Contact: yzhang{at}ku.edu Supplementary information: Supplementary data are availableat Bioinformatics online. Associate Editor: Anna Tramontano     

Exploring the effects of sparse restraints on protein structure prediction

One of the main barriers to accurate computational protein structure prediction is searching the vast space of protein conformations. Distance restraints or inter‐residue contacts have been used to reduce this search space, easing the discovery of the correct folded state. It has been suggested that about 1 contact for every 12 residues may be sufficient to predict structure at fold level accuracy. Here, we use coarse‐grained structure‐based models in conjunction with molecular dynamics simulations to examine this empirical prediction. We generate sparse contact maps for 15 proteins of varying sequence lengths and topologies and find that given perfect secondary‐structural information, a small fraction of the native contact map (5%‐10%) suffices to fold proteins to their correct native states. We also find that different sparse maps are not equivalent and we make several observations about the type of maps that are successful at such structure prediction. Long range contacts are found to encode more information than shorter range ones, especially for α and αβ‐proteins. However, this distinction reduces for β‐proteins. Choosing contacts that are a consensus from successful maps gives predictive sparse maps as does choosing contacts that are well spread out over the protein structure. Additionally, the folding of proteins can also be used to choose predictive sparse maps. Overall, we conclude that structure‐based models can be used to understand the efficacy of structure‐prediction restraints and could, in future, be tuned to include specific force‐field interactions, secondary structure errors and noise in the sparse maps.     

Can correct protein models be identified?

The ability to separate correct models of protein structures from less correct models is of the greatest importance for protein structure prediction methods. Several studies have examined the ability of different types of energy function to detect the native, or native-like, protein structure from a large set of decoys. In contrast to earlier studies, we examine here the ability to detect models that only show limited structural similarity to the native structure. These correct models are defined by the existence of a fragment that shows significant similarity between this model and the native structure. It has been shown that the existence of such fragments is useful for comparing the performance between different fold recognition methods and that this performance correlates well with performance in fold recognition. We have developed ProQ, a neural-network-based method to predict the quality of a protein model that extracts structural features, such as frequency of atom-atom contacts, and predicts the quality of a model, as measured either by LGscore or MaxSub. We show that ProQ performs at least as well as other measures when identifying the native structure and is better at the detection of correct models. This performance is maintained over several different test sets. ProQ can also be combined with the Pcons fold recognition predictor (Pmodeller) to increase its performance, with the main advantage being the elimination of a few high-scoring incorrect models. Pmodeller was successful in CASP5 and results from the latest LiveBench, LiveBench-6, indicating that Pmodeller has a higher specificity than Pcons alone.     

The roles of an ephrin and a semaphorin in patterning cell-cell contacts in C. elegans sensory organ development

Ephrins and semaphorins regulate a wide variety of developmental processes, including axon guidance and cell migration. We have studied the roles of the ephrin EFN-4 and the semaphorin MAB-20 in patterning cell-cell contacts among the cells that give rise to the ray sensory organs of Caenorhabditis elegans. In wild-type, contacts at adherens junctions form only between cells belonging to the same ray. In efn-4 and mab-20 mutants, ectopic contacts form between cells belonging to different rays. Ectopic contacts also occur in mutants in regulatory genes that specify ray morphological identity. We used efn-4 and mab-20 reporters to investigate whether these ray identity genes function through activating expression of efn-4 or mab-20 in ray cells. mab-20 reporter expression in ray cells was unaffected by mutants in the Pax6 homolog mab-18 and the Hox genes egl-5 and mab-5, suggesting that these genes do not regulate mab-20 expression. We find that mab-18 is necessary for activating efn-4 reporter expression, but this activity alone is not sufficient to account for mab-18 function in controlling cell-cell contact formation. In egl-5 mutants, efn-4 reporter expression in certain ray cells was increased, inconsistent with a simple repulsion model for efn-4 action. The evidence indicates that ray identity genes primarily regulate ray morphogenesis by pathways other than through regulation of expression of semaphorin and ephrin.     

Analysis of distance-based protein structure prediction by deep learning in CASP13

This paper reports the CASP13 results of distance-based contact prediction, threading, and folding methods implemented in three RaptorX servers, which are built upon the powerful deep convolutional residual neural network (ResNet) method initiated by us for contact prediction in CASP12. On the 32 CASP13 FM (free-modeling) targets with a median multiple sequence alignment (MSA) depth of 36, RaptorX yielded the best contact prediction among 46 groups and almost the best 3D structure modeling among all server groups without time-consuming conformation sampling. In particular, RaptorX achieved top L/5, L/2, and L long-range contact precision of 70%, 58%, and 45%, respectively, and predicted correct folds (TMscore > 0.5) for 18 of 32 targets. Further, RaptorX predicted correct folds for all FM targets with >300 residues (T0950-D1, T0969-D1, and T1000-D2) and generated the best 3D models for T0950-D1 and T0969-D1 among all groups. This CASP13 test confirms our previous findings: (a) predicted distance is more useful than contacts for both template-based and free modeling; and (b) structure modeling may be improved by integrating template and coevolutionary information via deep learning. This paper will discuss progress we have made since CASP12, the strength and weakness of our methods, and why deep learning performed much better in CASP13.     

On the optimal folding of protein molecules

The process of formation of a globular structure by a long molecular chain has been examined. In this process, various regions of the chain interact with one another. We classify the contacts thus formed as “correct” and “erroneous” ones. The correct contacts are those characteristic of the final native globular structure. All other contacts can be treated as erroneous. It is demonstrated that globule formation may proceed actually without formation and subsequent decay of erroneous contacts. Our model permits avoiding examination of numerous erroneous variants inasmuch as the regions of the chain that form correct contacts enter “long-range” interactions that at the same time can be highly selective. The existence of interactions of this kind facilitates the mutual approach and interaction of just those regions of the chain that yield correct contacts. Based on database analysis, it is shown that the model is valid not only for abstract structures but also for real polypeptide chains capable of forming protein globules and helical fibrils.     

Use of covariance analysis for the prediction of structural domain boundaries from multiple protein sequence alignments

Current methods for identification of domains within protein sequences require either structural information or the identification of homologous domain sequences in different sequence contexts. Knowledge of structural domain boundaries is important for fold recognition experiments and structural determination by X-ray crystallography or nuclear magnetic resonance spectroscopy using the divide-and-conquer approach. Here, a new and conceptually simple method for the identification of structural domain boundaries in multiple protein sequence alignments is presented. Analysis of covariance at positions within the alignment is first used to predict 3D contacts. By the nature of the domain as an independent folding unit, inter-domain predicted contacts are fewer than intra-domain predicted contacts. By analysing all possible domain boundaries and constructing a smoothed profile of predicted contact density (PCD), true structural domain boundaries are predicted as local profile minima associated with low PCD. A training data set is constructed from 52 non-homologous two-domain protein sequences of known 3D structure and used to determine optimal parameters for the profile analysis. The alignments in the training data set contained 48 +/- 17 (mean +/- SD) sequences and lengths of 257 +/- 121 residues. Of the 47 alignments yielding predictions, 35% of true domain boundaries are predicted to within 15 amino acids by the local profile minimum with the lowest profile value. Including predictions from the second- and third-lowest local minima increases the correct domain boundary coverage to 60%, whereas the lowest five local minima cover 79% of correct domain boundaries. Through further profile analysis, criteria are presented which reliably identify subsets of more accurate predictions. Retrospective analysis of CASP3 targets shows predictions of sufficient accuracy to enable dramatically improved fold recognition results. Finally, a prediction is made for geminivirus AL1 protein which is in full agreement with biochemical data, yielding a plausible, novel threading result.     

A conformation ensemble approach to protein residue-residue contact

Background

Protein residue-residue contact prediction is important for protein model generation and model evaluation. Here we develop a conformation ensemble approach to improve residue-residue contact prediction. We collect a number of structural models stemming from a variety of methods and implementations. The various models capture slightly different conformations and contain complementary information which can be pooled together to capture recurrent, and therefore more likely, residue-residue contacts.

Results

We applied our conformation ensemble approach to free modeling targets from both CASP8 and CASP9. Given a diverse ensemble of models, the method is able to achieve accuracies of. 48 for the top L/5 medium range contacts and. 36 for the top L/5 long range contacts for CASP8 targets (L being the target domain length). When applied to targets from CASP9, the accuracies of the top L/5 medium and long range contact predictions were. 34 and. 30 respectively.

Conclusions

When operating on a moderately diverse ensemble of models, the conformation ensemble approach is an effective means to identify medium and long range residue-residue contacts. An immediate benefit of the method is that when tied with a scoring scheme, it can be used to successfully rank models.     

GAPDOCK: a Genetic Algorithm Approach to Protein Docking in CAPRI round 1

As part of the first Critical Assessment of PRotein Interactions, round 1, we predict the structure of two protein-protein complexes, by using a genetic algorithm, GAPDOCK, in combination with surface complementarity, buried surface area, biochemical information, and human intervention. Among the five models submitted for target 1, HPr phosphocarrier protein (B. subtilis) and the hexameric HPr kinase (L. lactis), the best correctly predicts 17 of 52 interprotein contacts, whereas for target 2, bovine rotavirus VP6 protein-monoclonal antibody, the best model predicts 27 of 52 correct contacts. Given the difficult nature of the targets, these predictions are very encouraging and compare well with those obtained by other methods. Nevertheless, it is clear that there is a need for improved methods for distinguishing between "correct" and "plausible but incorrect" complexes.