Impact of AlphaFold on structure prediction of protein complexes: The CASP15-CAPRI experiment

We present the results for CAPRI Round 54, the 5th joint CASP-CAPRI protein assembly prediction challenge. The Round offered 37 targets, including 14 homodimers, 3 homo-trimers, 13 heterodimers including 3 antibody-antigen complexes, and 7 large assemblies. On average ~70 CASP and CAPRI predictor groups, including more than 20 automatics servers, submitted models for each target. A total of 21 941 models submitted by these groups and by 15 CAPRI scorer groups were evaluated using the CAPRI model quality measures and the DockQ score consolidating these measures. The prediction performance was quantified by a weighted score based on the number of models of acceptable quality or higher submitted by each group among their five best models. Results show substantial progress achieved across a significant fraction of the 60+ participating groups. High-quality models were produced for about 40% of the targets compared to 8% two years earlier. This remarkable improvement is due to the wide use of the AlphaFold2 and AlphaFold2-Multimer software and the confidence metrics they provide. Notably, expanded sampling of candidate solutions by manipulating these deep learning inference engines, enriching multiple sequence alignments, or integration of advanced modeling tools, enabled top performing groups to exceed the performance of a standard AlphaFold2-Multimer version used as a yard stick. This notwithstanding, performance remained poor for complexes with antibodies and nanobodies, where evolutionary relationships between the binding partners are lacking, and for complexes featuring conformational flexibility, clearly indicating that the prediction of protein complexes remains a challenging problem.

Critical Assessment of Methods for Predicting the 3D Structure of Proteins and Protein Complexes

Advances in a scientific discipline are often measured by small, incremental steps. In this review, we report on two intertwined disciplines in the protein structure prediction field, modeling of single chains and modeling of complexes, that have over decades emulated this pattern, as monitored by the community-wide blind prediction experiments CASP and CAPRI. However, over the past few years, dramatic advances were observed for the accurate prediction of single protein chains, driven by a surge of deep learning methodologies entering the prediction field. We review the mainscientific developments that enabled these recent breakthroughs and feature the important role of blind prediction experiments in building up and nurturing the structure prediction field. We discuss how the new wave of artificial intelligence-based methods is impacting the fields of computational and experimental structural biology and highlight areas in which deep learning methods are likely to lead to future developments, provided that major challenges are overcome.

Prediction of protein assemblies, the next frontier: The CASP14-CAPRI experiment

We present the results for CAPRI Round 50, the fourth joint CASP-CAPRI protein assembly prediction challenge. The Round comprised a total of twelve targets, including six dimers, three trimers, and three higher-order oligomers. Four of these were easy targets, for which good structural templates were available either for the full assembly, or for the main interfaces (of the higher-order oligomers). Eight were difficult targets for which only distantly related templates were found for the individual subunits. Twenty-five CAPRI groups including eight automatic servers submitted ~1250 models per target. Twenty groups including six servers participated in the CAPRI scoring challenge submitted ~190 models per target. The accuracy of the predicted models was evaluated using the classical CAPRI criteria. The prediction performance was measured by a weighted scoring scheme that takes into account the number of models of acceptable quality or higher submitted by each group as part of their five top-ranking models. Compared to the previous CASP-CAPRI challenge, top performing groups submitted such models for a larger fraction (70-75%) of the targets in this Round, but fewer of these models were of high accuracy. Scorer groups achieved stronger performance with more groups submitting correct models for 70-80% of the targets or achieving high accuracy predictions. Servers performed less well in general, except for the MDOCKPP and LZERD servers, who performed on par with human groups. In addition to these results, major advances in methodology are discussed, providing an informative overview of where the prediction of protein assemblies currently stands.

Insight into membraneless organelles and their associated proteins: Drivers, Clients and Regulators

In recent years, attention has been devoted to proteins forming immiscible liquid phases within the liquid intracellular medium, commonly referred to as membraneless organelles (MLO). These organelles enable the spatiotemporal associations of cellular components that exchange dynamically with the cellular milieu. The dysregulation of these liquid-liquid phase separation processes (LLPS) may cause various diseases including neurodegenerative pathologies and cancer, among others. Until very recently, databases containing information on proteins forming MLOs, as well as tools and resources facilitating their analysis, were missing. This has recently changed with the publication of 4 databases that focus on different types of experiments, sets of proteins, inclusion criteria, and levels of annotation or curation. In this study we integrate and analyze the information across these databases, complement their records, and produce a consolidated set of proteins that enables the investigation of the LLPS phenomenon. To gain insight into the features that characterize different types of MLOs and the roles of their associated proteins, they were grouped into categories: High Confidence MLO associated (including Drivers and reviewed proteins), Potential Clients and Regulators, according to their annotated functions. We show that none of the databases taken alone covers the data sufficiently to enable meaningful analysis, validating our integration effort as essential for gaining better understanding of phase separation and laying the foundations for the discovery of new proteins potentially involved in this important cellular process. Lastly, we developed a server, enabling customized selections of different sets of proteins based on MLO location, database, disorder content, among other attributes (https://forti.shinyapps.io/mlos/).

The impact of structural bioinformatics tools and resources on SARS-CoV-2 research and therapeutic strategies

SARS-CoV-2 is the causative agent of COVID-19, the ongoing global pandemic. It has posed a worldwide challenge to human health as no effective treatment is currently available to combat the disease. Its severity has led to unprecedented collaborative initiatives for therapeutic solutions against COVID-19. Studies resorting to structure-based drug design for COVID-19 are plethoric and show good promise. Structural biology provides key insights into 3D structures, critical residues/mutations in SARS-CoV-2 proteins, implicated in infectivity, molecular recognition and susceptibility to a broad range of host species. The detailed understanding of viral proteins and their complexes with host receptors and candidate epitope/lead compounds is the key to developing a structure-guided therapeutic design. Since the discovery of SARS-CoV-2, several structures of its proteins have been determined experimentally at an unprecedented speed and deposited in the Protein Data Bank. Further, specialized structural bioinformatics tools and resources have been developed for theoretical models, data on protein dynamics from computer simulations, impact of variants/mutations and molecular therapeutics. Here, we provide an overview of ongoing efforts on developing structural bioinformatics tools and resources for COVID-19 research. We also discuss the impact of these resources and structure-based studies, to understand various aspects of SARS-CoV-2 infection and therapeutic development. These include (i) understanding differences between SARS-CoV-2 and SARS-CoV, leading to increased infectivity of SARS-CoV-2, (ii) deciphering key residues in the SARS-CoV-2 involved in receptor-antibody recognition, (iii) analysis of variants in host proteins that affect host susceptibility to infection and (iv) analyses facilitating structure-based drug and vaccine design against SARS-CoV-2.

Integration of Data from Liquid-Liquid Phase Separation Databases Highlights Concentration and Dosage Sensitivity of LLPS Drivers

Liquid–liquid phase separation (LLPS) is a molecular process that leads to the formation of membraneless organelles, representing functionally specialized liquid-like cellular condensates formed by proteins and nucleic acids. Integrating the data on LLPS-associated proteins from dedicated databases revealed only modest agreement between them and yielded a high-confidence dataset of 89 human LLPS drivers. Analysis of the supporting evidence for our dataset uncovered a systematic and potentially concerning difference between protein concentrations used in a good fraction of the in vitro LLPS experiments, a key parameter that governs the phase behavior, and the proteomics-derived cellular abundance levels of the corresponding proteins. Closer scrutiny of the underlying experimental data enabled us to offer a sound rationale for this systematic difference, which draws on our current understanding of the cellular organization of the proteome and the LLPS process. In support of this rationale, we find that genes coding for our human LLPS drivers tend to be dosage-sensitive, suggesting that their cellular availability is tightly regulated to preserve their functional role in direct or indirect relation to condensate formation. Our analysis offers guideposts for increasing agreement between in vitro and in vivo studies, probing the roles of proteins in LLPS.

Distance-Based Metrics for Comparing Conformational Ensembles of Intrinsically Disordered Proteins

Intrinsically disordered proteins are proteins whose native functional states represent ensembles of highly diverse conformations. Such ensembles are a challenge for quantitative structure comparisons because their conformational diversity precludes optimal superimposition of the atomic coordinates necessary for deriving common similarity measures such as the root mean-square deviation of these coordinates. Here, we introduce superimposition-free metrics that are based on computing matrices of the Cα-Cα distance distributions within ensembles and comparing these matrices between ensembles. Differences between two matrices yield information on the similarity between specific regions of the polypeptide, whereas the global structural similarity is captured by the root mean-square difference between the medians of the Cα-Cα distance distributions of two ensembles. Together, our metrics enable rigorous investigations of structure-function relationships in conformational ensembles of intrinsically disordered proteins derived using experimental restraints or by molecular simulations and for proteins containing both structured and disordered regions.

Blind prediction of homo- and hetero-protein complexes: The CASP13-CAPRI experiment

We present the results for CAPRI Round 46, the third joint CASP-CAPRI protein assembly prediction challenge. The Round comprised a total of 20 targets including 14 homo-oligomers and 6 heterocomplexes. Eight of the homo-oligomer targets and one heterodimer comprised proteins that could be readily modeled using templates from the Protein Data Bank, often available for the full assembly. The remaining 11 targets comprised 5 homodimers, 3 heterodimers, and two higher-order assemblies. These were more difficult to model, as their prediction mainly involved "ab-initio" docking of subunit models derived from distantly related templates. A total of ~30 CAPRI groups, including 9 automatic servers, submitted on average ~2000 models per target. About 17 groups participated in the CAPRI scoring rounds, offered for most targets, submitting ~170 models per target. The prediction performance, measured by the fraction of models of acceptable quality or higher submitted across all predictors groups, was very good to excellent for the nine easy targets. Poorer performance was achieved by predictors for the 11 difficult targets, with medium and high quality models submitted for only 3 of these targets. A similar performance "gap" was displayed by scorer groups, highlighting yet again the unmet challenge of modeling the conformational changes of the protein components that occur upon binding or that must be accounted for in template-based modeling. Our analysis also indicates that residues in binding interfaces were less well predicted in this set of targets than in previous Rounds, providing useful insights for directions of future improvements.

Interplay of self-association and conformational flexibility in regulating protein function

Many functional roles have been attributed to homodimers, the most common mode of protein self-association, notably in the regulation of enzymes, ion channels, transporters and transcription factors. Here we review findings that offer new insights into the different roles conformational flexibility plays in regulating homodimer function. Intertwined homodimers of two-domain proteins and their related family members display significant conformational flexibility, which translates into concerted motion between structural domains. This flexibility enables the corresponding proteins to regulate function across family members by modulating the spatial positions of key recognition surfaces of individual domains, to either maintain subunit interfaces, alter them or break them altogether, leading to a variety of functional consequences. Many proteins may exist as monomers but carry out their biological function as homodimers or higher-order oligomers. We present early evidence that in such systems homodimer formation primes the protein for its functional role. It does so by inducing elevated mobility in protein regions corresponding to the binding epitopes of functionally important ligands. In some systems this process acts as an allosteric response elicited by the self-association reaction itself. Our analysis furthermore suggests that the induced extra mobility likely facilitates ligand binding through the mechanism of conformational selection.This article is part of a discussion meeting issue 'Allostery and molecular machines'.

The Balancing Act of Intrinsically Disordered Proteins: Enabling Functional Diversity while Minimizing Promiscuity

Intrinsically disordered proteins (IDPs) or regions (IDRs) perform diverse cellular functions, but are also prone to forming promiscuous and potentially deleterious interactions. We investigate the extent to which the properties of, and content in, IDRs have adapted to enable functional diversity while limiting interference from promiscuous interactions in the crowded cellular environment. Information on protein sequences, their predicted intrinsic disorder, and 3D structure contents is related to data on protein cellular concentrations, gene co-expression, and protein-protein interactions in the well-studied yeast Saccharomyces cerevisiae. Results reveal that both the protein IDR content and the frequency of "sticky" amino acids in IDRs (those more frequently involved in protein interfaces) decrease with increasing protein cellular concentration. This implies that the IDR content and the amino acid composition of IDRs experience negative selection as the protein concentration increases. In the S. cerevisiae protein-protein interaction network, the higher a protein's IDR content, the more frequently it interacts with IDR-containing partners, and the more functionally diverse the partners are. Employing a clustering analysis of Gene Ontology terms, we newly identify ~600 putative multifunctional proteins in S. cerevisiae. Strikingly, these proteins are enriched in IDRs and contribute significantly to all the observed trends. In particular, IDRs of multi-functional proteins feature more sticky amino acids than IDRs of their non-multifunctional counterparts, or the surfaces of structured yeast proteins. This property likely affords sufficient binding affinity for the functional interactions, commonly mediated by short IDR segments, thereby counterbalancing the loss in overall IDR conformational entropy upon binding.

The challenge of modeling protein assemblies: the CASP12-CAPRI experiment

We present the quality assessment of 5613 models submitted by predictor groups from both CAPRI and CASP for the total of 15 most tractable targets from the second joint CASP-CAPRI protein assembly prediction experiment. These targets comprised 12 homo-oligomers and 3 hetero-complexes. The bulk of the analysis focuses on 10 targets (of CAPRI Round 37), which included all 3 hetero-complexes, and whose protein chains or the full assembly could be readily modeled from structural templates in the PDB. On average, 28 CAPRI groups and 10 CASP groups (including automatic servers), submitted models for each of these 10 targets. Additionally, about 16 groups participated in the CAPRI scoring experiments. A range of acceptable to high quality models were obtained for 6 of the 10 Round 37 targets, for which templates were available for the full assembly. Poorer results were achieved for the remaining targets due to the lower quality of the templates available for the full complex or the individual protein chains, highlighting the unmet challenge of modeling the structural adjustments of the protein components that occur upon binding or which must be accounted for in template-based modeling. On the other hand, our analysis indicated that residues in binding interfaces were correctly predicted in a sizable fraction of otherwise poorly modeled assemblies and this with higher accuracy than published methods that do not use information on the binding partner. Lastly, the strengths and weaknesses of the assessment methods are evaluated and improvements suggested.

Interplay of buried histidine protonation and protein stability in prion misfolding

Misofolding of mammalian prion proteins (PrP) is believed to be the cause of a group of rare and fatal neurodegenerative diseases. Despite intense scrutiny however, the mechanism of the misfolding reaction remains unclear. We perform nuclear Magnetic Resonance and thermodynamic stability measurements on the C-terminal domains (residues 90–231) of two PrP variants exhibiting different pH-induced susceptibilities to aggregation: the susceptible hamster prion (GHaPrP) and its less susceptible rabbit homolog (RaPrP). The pKa of histidines in these domains are determined from titration experiments, and proton-exchange rates are measured at pH 5 and pH 7. A single buried highly conserved histidine, H187/H186 in GHaPrP/RaPrP, exhibited a markedly down shifted pKa ~5 for both proteins. However, noticeably larger pH-induced shifts in exchange rates occur for GHaPrP versus RaPrP. Analysis of the data indicates that protonation of the buried histidine destabilizes both PrP variants, but produces a more drastic effect in the less stable GHaPrP. This interpretation is supported by urea denaturation experiments performed on both PrP variants at neutral and low pH, and correlates with the difference in disease susceptibility of the two species, as expected from the documented linkage between destabilization of the folded state and formation of misfolded and aggregated species.

A benchmark testing ground for integrating homology modeling and protein docking

Protein docking procedures carry out the task of predicting the structure of a protein-protein complex starting from the known structures of the individual protein components. More often than not, however, the structure of one or both components is not known, but can be derived by homology modeling on the basis of known structures of related proteins deposited in the Protein Data Bank (PDB). Thus, the problem is to develop methods that optimally integrate homology modeling and docking with the goal of predicting the structure of a complex directly from the amino acid sequences of its component proteins. One possibility is to use the best available homology modeling and docking methods. However, the models built for the individual subunits often differ to a significant degree from the bound conformation in the complex, often much more so than the differences observed between free and bound structures of the same protein, and therefore additional conformational adjustments, both at the backbone and side chain levels need to be modeled to achieve an accurate docking prediction. In particular, even homology models of overall good accuracy frequently include localized errors that unfavorably impact docking results. The predicted reliability of the different regions in the model can also serve as a useful input for the docking calculations. Here we present a benchmark dataset that should help to explore and solve combined modeling and docking problems. This dataset comprises a subset of the experimentally solved 'target' complexes from the widely used Docking Benchmark from the Weng Lab (excluding antibody-antigen complexes). This subset is extended to include the structures from the PDB related to those of the individual components of each complex, and hence represent potential templates for investigating and benchmarking integrated homology modeling and docking approaches. Template sets can be dynamically customized by specifying ranges in sequence similarity and in PDB release dates, or using other filtering options, such as excluding sets of specific structures from the template list. Multiple sequence alignments, as well as structural alignments of the templates to their corresponding subunits in the target are also provided. The resource is accessible online or can be downloaded at http://cluspro.org/benchmark, and is updated on a weekly basis in synchrony with new PDB releases. Proteins 2016; 85:10-16. © 2016 Wiley Periodicals, Inc.

Computational proteome-wide screening predicts neurotoxic drug-protein interactome for the investigational analgesic BIA 10-2474

The investigational compound BIA 10-2474, designed as a long-acting and reversible inhibitor of fatty acid amide hydrolase for the treatment of neuropathic pain, led to the death of one participant and hospitalization of five others due to intracranial hemorrhage in a Phase I clinical trial. Putative off-target activities of BIA 10-2474 have been suggested to be major contributing factors to the observed neurotoxicity in humans, motivating our study's proteome-wide screening approach to investigate its polypharmacology. Accordingly, we performed an in silico screen against 80,923 protein structures reported in the Protein Data Bank. The resulting list of 284 unique human interactors was further refined using target-disease association analyses to a subset of proteins previously linked to neurological, intracranial, inflammatory, hemorrhagic or clotting processes and/or diseases. Eleven proteins were identified as potential targets of BIA 10-2474, and the two highest-scoring proteins, Factor VII and thrombin, both essential blood-clotting factors, were predicted to be inhibited by BIA 10-2474 and suggest a plausible mechanism of toxicity. Once this small molecule becomes commercially available, future studies will be conducted to evaluate the predicted inhibitory effect of BIA 10-2474 on blood clot formation specifically in the brain.

Modeling protein-protein and protein-peptide complexes: CAPRI 6th edition

We present the sixth report evaluating the performance of methods for predicting the atomic resolution structures of protein complexes offered as targets to the community-wide initiative on the Critical Assessment of Predicted Interactions (CAPRI). The evaluation is based on a total of 20,670 predicted models for 8 protein-peptide complexes, a novel category of targets in CAPRI, and 12 protein-protein targets in CAPRI prediction Rounds held during the years 2013-2016. For two of the protein-protein targets, the focus was on the prediction of side-chain conformation and positions of interfacial water molecules. Seven of the protein-protein targets were particularly challenging owing to their multicomponent nature, to conformational changes at the binding site, or to a combination of both. Encouragingly, the very large multiprotein complex with the nucleosome was correctly predicted, and correct models were submitted for the protein-peptide targets, but not for some of the challenging protein-protein targets. Models of acceptable quality or better were obtained for 14 of the 20 targets, including medium quality models for 13 targets and high quality models for 8 targets, indicating tangible progress of present-day computational methods in modeling protein complexes with increased accuracy. Our evaluation suggests that the progress stems from better integration of different modeling tools with docking procedures, as well as the use of more sophisticated evolutionary information to score models. Nonetheless, adequate modeling of conformational flexibility in interacting proteins remains an important area with a crucial need for improvement. Proteins 2017; 85:359-377. © 2016 Wiley Periodicals, Inc.

Computational approaches to investigating allostery

Allosteric regulation plays a key role in many biological processes, such as signal transduction, transcriptional regulation, and many more. It is rooted in fundamental thermodynamic and dynamic properties of macromolecular systems that are still poorly understood and are moreover modulated by the cellular context. Here we review the computational approaches used in the investigation of allosteric processes in protein systems. We outline how the models of allostery have evolved from their initial formulation in the sixties to the current views, which more fully account for the roles of the thermodynamic and dynamic properties of the system. We then describe the major classes of computational approaches employed to elucidate the mechanisms of allostery, the insights they have provided, as well as their limitations. We complement this analysis by highlighting the role of computational approaches in promising practical applications, such as the engineering of regulatory modules and identifying allosteric binding sites.

Prediction of homoprotein and heteroprotein complexes by protein docking and template-based modeling: A CASP-CAPRI experiment

We present the results for CAPRI Round 30, the first joint CASP-CAPRI experiment, which brought together experts from the protein structure prediction and protein-protein docking communities. The Round comprised 25 targets from amongst those submitted for the CASP11 prediction experiment of 2014. The targets included mostly homodimers, a few homotetramers, and two heterodimers, and comprised protein chains that could readily be modeled using templates from the Protein Data Bank. On average 24 CAPRI groups and 7 CASP groups submitted docking predictions for each target, and 12 CAPRI groups per target participated in the CAPRI scoring experiment. In total more than 9500 models were assessed against the 3D structures of the corresponding target complexes. Results show that the prediction of homodimer assemblies by homology modeling techniques and docking calculations is quite successful for targets featuring large enough subunit interfaces to represent stable associations. Targets with ambiguous or inaccurate oligomeric state assignments, often featuring crystal contact-sized interfaces, represented a confounding factor. For those, a much poorer prediction performance was achieved, while nonetheless often providing helpful clues on the correct oligomeric state of the protein. The prediction performance was very poor for genuine tetrameric targets, where the inaccuracy of the homology-built subunit models and the smaller pair-wise interfaces severely limited the ability to derive the correct assembly mode. Our analysis also shows that docking procedures tend to perform better than standard homology modeling techniques and that highly accurate models of the protein components are not always required to identify their association modes with acceptable accuracy. Proteins 2016; 84(Suppl 1):323-348. © 2016 Wiley Periodicals, Inc.

The Landscape of Intertwined Associations in Homooligomeric Proteins

We present an overview of the full repertoire of intertwined associations in homooligomeric proteins. This overview summarizes recent findings on the different categories of intertwined associations in known protein structures, their assembly modes, the properties of their interfaces, and their structural plasticity. Furthermore, the current body of knowledge on the so-called three-dimensional domain-swapped systems is reexamined in the context of the wider landscape of intertwined homooligomers, with a particular focus on the mechanistic aspects that underpin intertwined self-association processes in proteins. Insights gained from this integrated overview into the physical and biological roles of intertwining are highlighted.

Landscape of intertwined associations in multi-domain homo-oligomeric proteins

This study charts the landscape of multi-domain protein structures that form intertwined homodimers by exchanging structural domains between subunits. A representative dataset of such homodimers was derived from the Protein Data Bank, and their structural and topological properties were compared to those of a representative set of non-intertwined homodimers. Most of the intertwined dimers form closed assemblies with head-to-tail arrangements, where the subunit interface involves contacts between dissimilar domains. In contrast, the non-intertwined dimers form preferentially head-to-head arrangements, where the subunit interface involves contacts between identical domains. Most of these contacts engage only one structural domain from each subunit, leaving the remaining domains free to form other associations. Remarkably, we find that multi-domain proteins closely related to the intertwined homodimers are significantly more likely than relatives of the non-intertwined versions to adopt alternative intramolecular domain arrangements. In ~40% of the intertwined dimers, the plasticity in domain arrangements among relatives affords maintenance of the head-to-head or head-to-tail topology and conservation of the corresponding subunit interface. This property seems to be exploited in several systems to regulate DNA binding. In ~58%, however, intramolecular domain re-arrangements are associated with changes in oligomeric states and poorly conserved interfaces among relatives. This time, the corresponding structural plasticity appears to be exploited by evolution to modulate function by switching between active and inactive states of the protein. Surprisingly, in total, only three systems were found to undergo the classical monomer to intertwined dimer conversion associated with three-dimensional domain swapping.

Intercellular network structure and regulatory motifs in the human hematopoietic system

The hematopoietic system is a distributed tissue that consists of functionally distinct cell types continuously produced through hematopoietic stem cell (HSC) differentiation. Combining genomic and phenotypic data with high-content experiments, we have built a directional cell-cell communication network between 12 cell types isolated from human umbilical cord blood. Network structure analysis revealed that ligand production is cell type dependent, whereas ligand binding is promiscuous. Consequently, additional control strategies such as cell frequency modulation and compartmentalization were needed to achieve specificity in HSC fate regulation. Incorporating the in vitro effects (quiescence, self-renewal, proliferation, or differentiation) of 27 HSC binding ligands into the topology of the cell-cell communication network allowed coding of cell type-dependent feedback regulation of HSC fate. Pathway enrichment analysis identified intracellular regulatory motifs enriched in these cell type- and ligand-coupled responses. This study uncovers cellular mechanisms of hematopoietic cell feedback in HSC fate regulation, provides insight into the design principles of the human hematopoietic system, and serves as a foundation for the analysis of intercellular regulation in multicellular systems.

Blind prediction of interfacial water positions in CAPRI

We report the first assessment of blind predictions of water positions at protein-protein interfaces, performed as part of the critical assessment of predicted interactions (CAPRI) community-wide experiment. Groups submitting docking predictions for the complex of the DNase domain of colicin E2 and Im2 immunity protein (CAPRI Target 47), were invited to predict the positions of interfacial water molecules using the method of their choice. The predictions-20 groups submitted a total of 195 models-were assessed by measuring the recall fraction of water-mediated protein contacts. Of the 176 high- or medium-quality docking models-a very good docking performance per se-only 44% had a recall fraction above 0.3, and a mere 6% above 0.5. The actual water positions were in general predicted to an accuracy level no better than 1.5 Å, and even in good models about half of the contacts represented false positives. This notwithstanding, three hotspot interface water positions were quite well predicted, and so was one of the water positions that is believed to stabilize the loop that confers specificity in these complexes. Overall the best interface water predictions was achieved by groups that also produced high-quality docking models, indicating that accurate modelling of the protein portion is a determinant factor. The use of established molecular mechanics force fields, coupled to sampling and optimization procedures also seemed to confer an advantage. Insights gained from this analysis should help improve the prediction of protein-water interactions and their role in stabilizing protein complexes.

Navigating the global protein-protein interaction landscape using iRefWeb

iRefWeb is a bioinformatics resource that offers access to a large collection of data on protein-protein interactions in over a thousand organisms. This collection is consolidated from 14 major public databases that curate the scientific literature. The collection is enhanced with a range of versatile data filters and search options that categorize various types of protein-protein interactions and protein complexes. Users of iRefWeb are able to retrieve all curated interactions for a given organism or those involving a given protein (or a list of proteins), narrow down their search results based on different supporting evidence, and assess the reliability of these interactions using various criteria. They may also examine all data and annotations related to any publication that described the interaction-detection experiments. iRefWeb is freely available to the research community worldwide at http://wodaklab.org/iRefWeb .

Community-wide evaluation of methods for predicting the effect of mutations on protein-protein interactions

Community-wide blind prediction experiments such as CAPRI and CASP provide an objective measure of the current state of predictive methodology. Here we describe a community-wide assessment of methods to predict the effects of mutations on protein-protein interactions. Twenty-two groups predicted the effects of comprehensive saturation mutagenesis for two designed influenza hemagglutinin binders and the results were compared with experimental yeast display enrichment data obtained using deep sequencing. The most successful methods explicitly considered the effects of mutation on monomer stability in addition to binding affinity, carried out explicit side-chain sampling and backbone relaxation, evaluated packing, electrostatic, and solvation effects, and correctly identified around a third of the beneficial mutations. Much room for improvement remains for even the best techniques, and large-scale fitness landscapes should continue to provide an excellent test bed for continued evaluation of both existing and new prediction methodologies.

A census of human soluble protein complexes

Cellular processes often depend on stable physical associations between proteins. Despite recent progress, knowledge of the composition of human protein complexes remains limited. To close this gap, we applied an integrative global proteomic profiling approach, based on chromatographic separation of cultured human cell extracts into more than one thousand biochemical fractions that were subsequently analyzed by quantitative tandem mass spectrometry, to systematically identify a network of 13,993 high-confidence physical interactions among 3,006 stably associated soluble human proteins. Most of the 622 putative protein complexes we report are linked to core biological processes and encompass both candidate disease genes and unannotated proteins to inform on mechanism. Strikingly, whereas larger multiprotein assemblies tend to be more extensively annotated and evolutionarily conserved, human protein complexes with five or fewer subunits are far more likely to be functionally unannotated or restricted to vertebrates, suggesting more recent functional innovations.

Interaction landscape of membrane-protein complexes in Saccharomyces cerevisiae

Macromolecular assemblies involving membrane proteins (MPs) serve vital biological roles and are prime drug targets in a variety of diseases. Large-scale affinity purification studies of soluble-protein complexes have been accomplished for diverse model organisms, but no global characterization of MP-complex membership has been described so far. Here we report a complete survey of 1,590 putative integral, peripheral and lipid-anchored MPs from Saccharomyces cerevisiae, which were affinity purified in the presence of non-denaturing detergents. The identities of the co-purifying proteins were determined by tandem mass spectrometry and subsequently used to derive a high-confidence physical interaction map encompassing 1,726 membrane protein-protein interactions and 501 putative heteromeric complexes associated with the various cellular membrane systems. Our analysis reveals unexpected physical associations underlying the membrane biology of eukaryotes and delineates the global topological landscape of the membrane interactome.

Savant Genome Browser 2: visualization and analysis for population-scale genomics

High-throughput sequencing (HTS) technologies are providing an unprecedented capacity for data generation, and there is a corresponding need for efficient data exploration and analysis capabilities. Although most existing tools for HTS data analysis are developed for either automated (e.g. genotyping) or visualization (e.g. genome browsing) purposes, such tools are most powerful when combined. For example, integration of visualization and computation allows users to iteratively refine their analyses by updating computational parameters within the visual framework in real-time. Here we introduce the second version of the Savant Genome Browser, a standalone program for visual and computational analysis of HTS data. Savant substantially improves upon its predecessor and existing tools by introducing innovative visualization modes and navigation interfaces for several genomic datatypes, and synergizing visual and automated analyses in a way that is powerful yet easy even for non-expert users. We also present a number of plugins that were developed by the Savant Community, which demonstrate the power of integrating visual and automated analyses using Savant. The Savant Genome Browser is freely available (open source) at www.savantbrowser.com.

Multiple replica repulsion technique for efficient conformational sampling of biological systems

Here, we propose a technique for sampling complex molecular systems with many degrees of freedom. The technique, termed "multiple replica repulsion" (MRR), does not suffer from poor scaling with the number of degrees of freedom associated with common replica exchange procedures and does not require sampling at high temperatures. The algorithm involves creation of multiple copies (replicas) of the system, which interact with one another through a repulsive potential that can be applied to the system as a whole or to portions of it. The proposed scheme prevents oversampling of the most populated states and provides accurate descriptions of conformational perturbations typically associated with sampling ground-state energy wells. The performance of MRR is illustrated for three systems of increasing complexity. A two-dimensional toy potential surface is used to probe the sampling efficiency as a function of key parameters of the procedure. MRR simulations of the Met-enkephalin pentapeptide, and the 76-residue protein ubiquitin, performed in presence of explicit water molecules and totaling 32 ns each, investigate the ability of MRR to characterize the conformational landscape of the peptide, and the protein native basin, respectively. Results obtained for the enkephalin peptide reflect more closely the extensive conformational flexibility of this peptide than previously reported simulations. Those obtained for ubiquitin show that conformational ensembles sampled by MRR largely encompass structural fluctuations relevant to biological recognition, which occur on the microsecond timescale, or are observed in crystal structures of ubiquitin complexes with other proteins. MRR thus emerges as a very promising simple and versatile technique for modeling the structural plasticity of complex biological systems.

Community-wide assessment of protein-interface modeling suggests improvements to design methodology

The CAPRI (Critical Assessment of Predicted Interactions) and CASP (Critical Assessment of protein Structure Prediction) experiments have demonstrated the power of community-wide tests of methodology in assessing the current state of the art and spurring progress in the very challenging areas of protein docking and structure prediction. We sought to bring the power of community-wide experiments to bear on a very challenging protein design problem that provides a complementary but equally fundamental test of current understanding of protein-binding thermodynamics. We have generated a number of designed protein-protein interfaces with very favorable computed binding energies but which do not appear to be formed in experiments, suggesting that there may be important physical chemistry missing in the energy calculations. A total of 28 research groups took up the challenge of determining what is missing: we provided structures of 87 designed complexes and 120 naturally occurring complexes and asked participants to identify energetic contributions and/or structural features that distinguish between the two sets. The community found that electrostatics and solvation terms partially distinguish the designs from the natural complexes, largely due to the nonpolar character of the designed interactions. Beyond this polarity difference, the community found that the designed binding surfaces were, on average, structurally less embedded in the designed monomers, suggesting that backbone conformational rigidity at the designed surface is important for realization of the designed function. These results can be used to improve computational design strategies, but there is still much to be learned; for example, one designed complex, which does form in experiments, was classified by all metrics as a nonbinder.

Assessment of methylation level prediction accuracy in methyl-DNA immunoprecipitation and sodium bisulfite based microarray platforms

In this study, we verified the accuracy of two array methods--methylated DNA immunoprecipitation coupled with CpG island microarrays (MeDIP-CGI-arrays) and sodium bisulfite conversion based microarrays (BC-arrays)--in predicting regional methylation levels as measured by pyrosequencing of bisulfite converted DNA (BC-pyrosequencing). To test the accuracy of these methods we used the Agilent Human CpG island and the Illumina HumanMethylation27 microarrays respectively, and compared microarray outputs to the data from targeted BC-pyrosequencing assays from several genomic regions of corresponding samples. We observed relatively high correlation with BC-pyrosequencing data for both array platforms, R = 0.87 for BC-Array and R = 0.79 for MeDIP-CGI array. However, MeDIP-CGI array were less reliable in predicting intermediate levels of DNA methylation. Several bioinformatics strategies, to ameliorate the performance of the MeDIP-CGI-Arrays did not improve the correlation with BC-pyrosequencing data. The high scalability, low cost and simpler analysis of BC-arrays, together with the recent extended coverage may make them a more versatile methylation analysis tool.

Docking and scoring protein interactions: CAPRI 2009

Protein docking algorithms are assessed by evaluating blind predictions performed during 2007-2009 in Rounds 13-19 of the community-wide experiment on critical assessment of predicted interactions (CAPRI). We evaluated the ability of these algorithms to sample docking poses and to single out specific association modes in 14 targets, representing 11 distinct protein complexes. These complexes play important biological roles in RNA maturation, G-protein signal processing, and enzyme inhibition and function. One target involved protein-RNA interactions not previously considered in CAPRI, several others were hetero-oligomers, or featured multiple interfaces between the same protein pair. For most targets, predictions started from the experimentally determined structures of the free (unbound) components, or from models built from known structures of related or similar proteins. To succeed they therefore needed to account for conformational changes and model inaccuracies. In total, 64 groups and 12 web-servers submitted docking predictions of which 4420 were evaluated. Overall our assessment reveals that 67% of the groups, more than ever before, produced acceptable models or better for at least one target, with many groups submitting multiple high- and medium-accuracy models for two to six targets. Forty-one groups including four web-servers participated in the scoring experiment with 1296 evaluated models. Scoring predictions also show signs of progress evidenced from the large proportion of correct models submitted. But singling out the best models remains a challenge, which also adversely affects the ability to correctly rank docking models. With the increased interest in translating abstract protein interaction networks into realistic models of protein assemblies, the growing CAPRI community is actively developing more efficient and reliable docking and scoring methods for everyone to use.

OrthoNets: simultaneous visual analysis of orthologs and their interaction neighborhoods across different organisms

Motivation: Protein interaction networks contain a wealth of biological information, but their large size often hinders cross-organism comparisons. We present OrthoNets, a Cytoscape plugin that displays protein–protein interaction (PPI) networks from two organisms simultaneously, highlighting orthology relationships and aggregating several types of biomedical annotations. OrthoNets also allows PPI networks derived from experiments to be overlaid on networks extracted from public databases, supporting the identification and verification of new interactors. Any newly identified PPIs can be validated by checking whether their orthologs interact in another organism.

Availability: OrthoNets is freely available at http://wodaklab.org/orthonets/.

Contact:jim.vlasblom@utoronto.ca

High-throughput analyses and curation of protein interactions in yeast

The yeast Saccharomyces cerevisiae is the model organism in which protein interactions have been most extensively analyzed. The vast majority of these interactions have been characterized by a variety of sophisticated high-throughput techniques probing different aspects of protein association. This chapter summarizes the major techniques, highlights their complementary nature, discusses the data they produce, and highlights some of the biases from which they suffer. A main focus is the key role played by computational methods for processing, analyzing, and validating the large body of noisy data produced by the experimental procedures. It also describes how computational methods are used to extend the coverage and reliability of protein interaction data by integrating information from heterogeneous sources and reviews the current status of literature-curated data on yeast protein interactions stored in specialized databases.

DAnCER: disease-annotated chromatin epigenetics resource

Chromatin modification (CM) is a set of epigenetic processes that govern many aspects of DNA replication, transcription and repair. CM is carried out by groups of physically interacting proteins, and their disruption has been linked to a number of complex human diseases. CM remains largely unexplored, however, especially in higher eukaryotes such as human. Here we present the DAnCER resource, which integrates information on genes with CM function from five model organisms, including human. Currently integrated are gene functional annotations, Pfam domain architecture, protein interaction networks and associated human diseases. Additional supporting evidence includes orthology relationships across organisms, membership in protein complexes, and information on protein 3D structure. These data are available for 962 experimentally confirmed and manually curated CM genes and for over 5000 genes with predicted CM function on the basis of orthology and domain composition. DAnCER allows visual explorations of the integrated data and flexible query capabilities using a variety of data filters. In particular, disease information and functional annotations are mapped onto the protein interaction networks, enabling the user to formulate new hypotheses on the function and disease associations of a given gene based on those of its interaction partners. DAnCER is freely available at http://wodaklab.org/dancer/.

Literature curation of protein interactions: measuring agreement across major public databases

Literature curation of protein interaction data faces a number of challenges. Although curators increasingly adhere to standard data representations, the data that various databases actually record from the same published information may differ significantly. Some of the reasons underlying these differences are well known, but their global impact on the interactions collectively curated by major public databases has not been evaluated. Here we quantify the agreement between curated interactions from 15 471 publications shared across nine major public databases. Results show that on average, two databases fully agree on 42% of the interactions and 62% of the proteins curated from the same publication. Furthermore, a sizable fraction of the measured differences can be attributed to divergent assignments of organism or splice isoforms, different organism focus and alternative representations of multi-protein complexes. Our findings highlight the impact of divergent curation policies across databases, and should be relevant to both curators and data consumers interested in analyzing protein-interaction data generated by the scientific community.

Database URL:http://wodaklab.org/iRefWeb