Assessing the utility of coevolution-based residue-residue contact predictions in a sequence- and structure-rich era

Structure determination of the HgcAB complex using metagenome sequence data: insights into microbial mercury methylation

Bacteria and archaea possessing the hgcAB gene pair methylate inorganic mercury (Hg) to form highly toxic methylmercury. HgcA consists of a corrinoid binding domain and a transmembrane domain, and HgcB is a dicluster ferredoxin. However, their detailed structure and function have not been thoroughly characterized. We modeled the HgcAB complex by combining metagenome sequence data mining, coevolution analysis, and Rosetta structure calculations. In addition, we overexpressed HgcA and HgcB in Escherichia coli, confirmed spectroscopically that they bind cobalamin and [4Fe-4S] clusters, respectively, and incorporated these cofactors into the structural model. Surprisingly, the two domains of HgcA do not interact with each other, but HgcB forms extensive contacts with both domains. The model suggests that conserved cysteines in HgcB are involved in shuttling Hg^II, methylmercury, or both. These findings refine our understanding of the mechanism of Hg methylation and expand the known repertoire of corrinoid methyltransferases in nature.

Connor J. Cooper et al. expressed HgcA and HgcB in Escherichia coli and modeled the structure of the HgcAB complex by combining metagenome sequence data, coevolution analysis, and ab initio structure calculations. This study provides insights into the biochemical mechanism of mercury (Hg) methylation.

Protein covariance networks reveal interactions important to the emergence of SARS coronaviruses as human pathogens

SARS-CoV-2 is one of three recognized coronaviruses (CoVs) that have caused epidemics or pandemics in the 21 ^st century and that have likely emerged from animal reservoirs based on genomic similarities to bat and other animal viruses. Here we report the analysis of conserved interactions between amino acid residues in proteins encoded by SARS-CoV-related viruses. We identified pairs and networks of residue variants that exhibited statistically high frequencies of covariance with each other. While these interactions are likely key to both protein structure and other protein-protein interactions, we have also found that they can be used to provide a new computational approach (CoVariance-based Phylogeny Analysis) for understanding viral evolution and adaptation. Our data provide evidence that the evolutionary processes that converted a bat virus into human pathogen occurred through recombination with other viruses in combination with new adaptive mutations important for entry into human cells.

Information Theory in Computational Biology: Where We Stand Today

"A Mathematical Theory of Communication" was published in 1948 by Claude Shannon to address the problems in the field of data compression and communication over (noisy) communication channels. Since then, the concepts and ideas developed in Shannon's work have formed the basis of information theory, a cornerstone of statistical learning and inference, and has been playing a key role in disciplines such as physics and thermodynamics, probability and statistics, computational sciences and biological sciences. In this article we review the basic information theory based concepts and describe their key applications in multiple major areas of research in computational biology-gene expression and transcriptomics, alignment-free sequence comparison, sequencing and error correction, genome-wide disease-gene association mapping, metabolic networks and metabolomics, and protein sequence, structure and interaction analysis.

GalaxyRefine2: simultaneous refinement of inaccurate local regions and overall protein structure

The 3D structure of a protein can be predicted from its amino acid sequence with high accuracy for a large fraction of cases because of the availability of large quantities of experimental data and the advance of computational algorithms. Recently, deep learning methods exploiting the coevolution information obtained by comparing related protein sequences have been successfully used to generate highly accurate model structures even in the absence of template structure information. However, structures predicted based on either template structures or related sequences require further improvement in regions for which information is missing. Refining a predicted protein structure with insufficient information on certain regions is critical because these regions may be connected to functional specificity that is not conserved among related proteins. The GalaxyRefine2 web server, freely available via http://galaxy.seoklab.org/refine2, is an upgraded version of the GalaxyRefine protein structure refinement server and reflects recent developments successfully tested through CASP blind prediction experiments. This method adopts an iterative optimization approach involving various structure move sets to refine both local and global structures. The estimation of local error and hybridization of available homolog structures are also employed for effective conformation search.

Coevolutionary data-based interaction networks approach highlighting key residues across protein families: The case of the G-protein coupled receptors

We present an approach that, by integrating structural data with Direct Coupling Analysis, is able to pinpoint most of the interaction hotspots (i.e. key residues for the biological activity) across very sparse protein families in a single run. An application to the Class A G-protein coupled receptors (GPCRs), both in their active and inactive states, demonstrates the predictive power of our approach. The latter can be easily extended to any other kind of protein family, where it is expected to highlight most key sites involved in their functional activity.

Multifaceted analysis of training and testing convolutional neural networks for protein secondary structure prediction

Protein secondary structure prediction remains a vital topic with broad applications. Due to lack of a widely accepted standard in secondary structure predictor evaluation, a fair comparison of predictors is challenging. A detailed examination of factors that contribute to higher accuracy is also lacking. In this paper, we present: (1) new test sets, Test2018, Test2019, and Test2018-2019, consisting of proteins from structures released in 2018 and 2019 with less than 25% identity to any protein published before 2018; (2) a 4-layer convolutional neural network, SecNet, with an input window of ±14 amino acids which was trained on proteins ≤25% identical to proteins in Test2018 and the commonly used CB513 test set; (3) an additional test set that shares no homologous domains with the training set proteins, according to the Evolutionary Classification of Proteins (ECOD) database; (4) a detailed ablation study where we reverse one algorithmic choice at a time in SecNet and evaluate the effect on the prediction accuracy; (5) new 4- and 5-label prediction alphabets that may be more practical for tertiary structure prediction methods. The 3-label accuracy (helix, sheet, coil) of the leading predictors on both Test2018 and CB513 is 81-82%, while SecNet's accuracy is 84% for both sets. Accuracy on the non-homologous ECOD set is only 0.6 points (83.9%) lower than the results on the Test2018-2019 set (84.5%). The ablation study of features, neural network architecture, and training hyper-parameters suggests the best accuracy results are achieved with good choices for each of them while the neural network architecture is not as critical as long as it is not too simple. Protocols for generating and using unbiased test, validation, and training sets are provided. Our data sets, including input features and assigned labels, and SecNet software including third-party dependencies and databases, are downloadable from dunbrack.fccc.edu/ss and github.com/sh-maxim/ss.

MAFFT online service: multiple sequence alignment, interactive sequence choice and visualization

This article describes several features in the MAFFT online service for multiple sequence alignment (MSA). As a result of recent advances in sequencing technologies, huge numbers of biological sequences are available and the need for MSAs with large numbers of sequences is increasing. To extract biologically relevant information from such data, sophistication of algorithms is necessary but not sufficient. Intuitive and interactive tools for experimental biologists to semiautomatically handle large data are becoming important. We are working on development of MAFFT toward these two directions. Here, we explain (i) the Web interface for recently developed options for large data and (ii) interactive usage to refine sequence data sets and MSAs.

Protein Contact Map Prediction Based on ResNet and DenseNet

Residue-residue contact prediction has become an increasingly important tool for modeling the three-dimensional structure of a protein when no homologous structure is available. Ultradeep residual neural network (ResNet) has become the most popular method for making contact predictions because it captures the contextual information between residues. In this paper, we propose a novel deep neural network framework for contact prediction which combines ResNet and DenseNet. This framework uses 1D ResNet to process sequential features, and besides PSSM, SS3, and solvent accessibility, we have introduced a new feature, position-specific frequency matrix (PSFM), as an input. Using ResNet's residual module and identity mapping, it can effectively process sequential features after which the outer concatenation function is used for sequential and pairwise features. Prediction accuracy is improved following a final processing step using the dense connection of DenseNet. The prediction accuracy of the protein contact map shows that our method is more effective than other popular methods due to the new network architecture and the added feature input.

Protein Structural Information and Evolutionary Landscape by In Vitro Evolution

Protein structure is tightly intertwined with function according to the laws of evolution. Understanding how structure determines function has been the aim of structural biology for decades. Here, we have wondered instead whether it is possible to exploit the function for which a protein was evolutionary selected to gain information on protein structure and on the landscape explored during the early stages of molecular and natural evolution. To answer to this question, we developed a new methodology, which we named CAMELS (Coupling Analysis by Molecular Evolution Library Sequencing), that is able to obtain the in vitro evolution of a protein from an artificial selection based on function. We were able to observe with CAMELS many features of the TEM-1 beta-lactamase local fold exclusively by generating and sequencing large libraries of mutational variants. We demonstrated that we can, whenever a functional phenotypic selection of a protein is available, sketch the structural and evolutionary landscape of a protein without utilizing purified proteins, collecting physical measurements, or relying on the pool of natural protein variants.

IMPContact: An Interhelical Residue Contact Prediction Method

As an important category of proteins, alpha-helix transmembrane proteins (αTMPs) play an important role in various biological activities. Because the solved αTMP structures are inadequate, predicting the residue contacts among the transmembrane segments of an αTMP exhibits the basis of protein fold, which can be used to further discover more protein functions. A few efforts have been devoted to predict the interhelical residue contact using machine learning methods based on the prior knowledge of transmembrane protein structure. However, it is still a challenge to improve the prediction accuracy, while the deep learning method provides an opportunity to utilize the structural knowledge in a different insight. For this purpose, we proposed a novel αTMP residue-residue contact prediction method IMPContact, in which a convolutional neural network (CNN) was applied to recognize those interhelical contacts in a TMP using its specific structural features. There were four sequence-based TMP-specific features selected to descript a pair of residues, namely, evolutionary covariation, predicted topology structure, residue relative position, and evolutionary conservation. An up-to-date dataset was used to train and test the IMPContact; our method achieved better performance compared to peer methods. In the case studies, IHRCs in the regular transmembrane helixes were better predicted than in the irregular ones.

Synthetic sequence entanglement augments stability and containment of genetic information in cells

In synthetic biology, methods for stabilizing genetically engineered functions and confining recombinant DNA to intended hosts are necessary to cope with natural mutation accumulation and pervasive lateral gene flow. We present a generalizable strategy to preserve and constrain genetic information through the computational design of overlapping genes. Overlapping a sequence with an essential gene altered its fitness landscape and produced a constrained evolutionary path, even for synonymous mutations. Embedding a toxin gene in a gene of interest restricted its horizontal propagation. We further demonstrated a multiplex and scalable approach to build and test >7500 overlapping sequence designs, yielding functional yet highly divergent variants from natural homologs. This work enables deeper exploration of natural and engineered overlapping genes and facilitates enhanced genetic stability and biocontainment in emerging applications.

Will Cryo-Electron Microscopy Shift the Current Paradigm in Protein Structure Prediction?

Protein dynamics is undoubtedly a pervasive ingredient in all biological functions. However, structural biology has been strongly driven by a static-centered view of protein architecture. We argue that the recent advances of cryo-electron microscopy (EM) have the potential to more broadly explore the conformational landscapes of protein complexes and therefore will enhance our ability to predict the diverse conformations of tertiary and quaternary protein structures that are functionally relevant in physiological conditions.

Improved Consensus-Fragment Selection in Template-Assisted Prediction of Protein Structures with the UNRES Force Field in CASP13

The method for protein-structure prediction, which combines the physics-based coarse-grained UNRES force field with knowledge-based modeling, has been developed further and tested in the 13th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP13). The method implements restraints from the consensus fragments common to server models. In this work, the server models to derive fragments have been chosen on the basis of quality assessment; a fully automatic fragment-selection procedure has been introduced, and Dynamic Fragment Assembly pseudopotentials have been fully implemented. The Global Distance Test Score (GDT_TS), averaged over our “Model 1” predictions, increased by over 10 units with respect to CASP12 for the free-modeling category to reach 40.82. Our “Model 1” predictions ranked 20 and 14 for all and free-modeling targets, respectively (upper 20.2% and 14.3% of all models submitted to CASP13 in these categories, respectively), compared to 27 (upper 21.1%) and 24 (upper 18.9%) in CASP12, respectively. For oligomeric targets, the Interface Patch Similarity (IPS) and Interface Contact Similarity (ICS) averaged over our best oligomer models increased from 0.28 to 0.36 and from 12.4 to 17.8, respectively, from CASP12 to CASP13, and top-ranking models of 2 targets (H0968 and T0997o) were obtained (none in CASP12). The improvement of our method in CASP13 over CASP12 was ascribed to the combined effect of the overall enhancement of server-model quality, our success in selecting server models and fragments to derive restraints, and improvements of the restraint and potential-energy functions.

DeepECA: an end-to-end learning framework for protein contact prediction from a multiple sequence alignment

Background

Recently developed methods of protein contact prediction, a crucially important step for protein structure prediction, depend heavily on deep neural networks (DNNs) and multiple sequence alignments (MSAs) of target proteins. Protein sequences are accumulating to an increasing degree such that abundant sequences to construct an MSA of a target protein are readily obtainable. Nevertheless, many cases present different ends of the number of sequences that can be included in an MSA used for contact prediction. The abundant sequences might degrade prediction results, but opportunities remain for a limited number of sequences to construct an MSA. To resolve these persistent issues, we strove to develop a novel framework using DNNs in an end-to-end manner for contact prediction.

Results

We developed neural network models to improve precision of both deep and shallow MSAs. Results show that higher prediction accuracy was achieved by assigning weights to sequences in a deep MSA. Moreover, for shallow MSAs, adding a few sequential features was useful to increase the prediction accuracy of long-range contacts in our model. Based on these models, we expanded our model to a multi-task model to achieve higher accuracy by incorporating predictions of secondary structures and solvent-accessible surface areas. Moreover, we demonstrated that ensemble averaging of our models can raise accuracy. Using past CASP target protein domains, we tested our models and demonstrated that our final model is superior to or equivalent to existing meta-predictors.

Conclusions

The end-to-end learning framework we built can use information derived from either deep or shallow MSAs for contact prediction. Recently, an increasing number of protein sequences have become accessible, including metagenomic sequences, which might degrade contact prediction results. Under such circumstances, our model can provide a means to reduce noise automatically. According to results of tertiary structure prediction based on contacts and secondary structures predicted by our model, more accurate three-dimensional models of a target protein are obtainable than those from existing ECA methods, starting from its MSA. DeepECA is available from https://github.com/tomiilab/DeepECA.

Improved protein structure prediction using predicted interresidue orientations

The prediction of interresidue contacts and distances from coevolutionary data using deep learning has considerably advanced protein structure prediction. Here, we build on these advances by developing a deep residual network for predicting interresidue orientations, in addition to distances, and a Rosetta-constrained energy-minimization protocol for rapidly and accurately generating structure models guided by these restraints. In benchmark tests on 13th Community-Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP13)- and Continuous Automated Model Evaluation (CAMEO)-derived sets, the method outperforms all previously described structure-prediction methods. Although trained entirely on native proteins, the network consistently assigns higher probability to de novo-designed proteins, identifying the key fold-determining residues and providing an independent quantitative measure of the "ideality" of a protein structure. The method promises to be useful for a broad range of protein structure prediction and design problems.

Coevolution analysis of amino-acids reveals diversified drug-resistance solutions in viral sequences: a case study of hepatitis B virus

The study of mutational landscapes of viral proteins is fundamental for the understanding of the mechanisms of cross-resistance to drugs and the design of effective therapeutic strategies based on several drugs. Antiviral therapy with nucleos(t)ide analogues targeting the hepatitis B virus (HBV) polymerase protein (Pol) can inhibit disease progression by suppression of HBV replication and makes it an important case study. In HBV, treatment may fail due to the emergence of drug-resistant mutants. Primary and compensatory mutations have been associated with lamivudine resistance, whereas more complex mutational patterns are responsible for resistance to other HBV antiviral drugs. So far, all known drug-resistance mutations are located in one of the four Pol domains, called reverse transcriptase. We demonstrate that sequence covariation identifies drug-resistance mutations in viral sequences. A new algorithmic strategy, BIS2TreeAnalyzer, is designed to apply the coevolution analysis method BIS2, successfully used in the past on small sets of conserved sequences, to large sets of evolutionary related sequences. When applied to HBV, BIS2TreeAnalyzer highlights diversified viral solutions by discovering thirty-seven positions coevolving with residues known to be associated with drug resistance and located on the four Pol domains. These results suggest a sequential mechanism of emergence for some mutational patterns. They reveal complex combinations of positions involved in HBV drug resistance and contribute with new information to the landscape of HBV evolutionary solutions. The computational approach is general and can be applied to other viral sequences when compensatory mutations are presumed.

A Benchmark for Homomeric Enzyme Active Site Structure Prediction Highlights the Importance of Accurate Modeling of Protein Symmetry

Accurate prediction and modeling of an enzyme's active site are critical for engineering efforts as well as providing insight into an enzyme's naturally occurring function. Previous efforts demonstrated that the integration of constraints enforcing strict geometric orientations between catalytic residues significantly improved the modeling accuracy for the active sites of monomeric enzymes. In this study, a similar approach was explored to evaluate the effect on the active sites of homomeric enzymes. A benchmark of 17 homomeric enzymes with known structures and a bound ligand relevant to the established chemistry were identified from the protein data bank. The enzymes identified span multiple classes as well as symmetries. Unlike what was observed for the monomeric enzymes, upon the application of catalytic geometric constraints, there was no significant improvement observed in modeling accuracy for either the active site of the protein structure or the accuracy of the subsequently docked ligand. Upon further analysis, it is apparent that the symmetric interface being modeled is inaccurate and prevented the active sites from being modeled at atomic-level accuracy. This is consistent with the challenge others have identified in being able to predict de novo protein symmetry. To further improve the accuracy of active site modeling for homomeric proteins, new methodologies to accurately model the symmetric interfaces of these complexes are needed.

Computational reconstruction of atomistic protein structures from coarse-grained models

Three-dimensional protein structures, whether determined experimentally or theoretically, are often too low resolution. In this mini-review, we outline the computational methods for protein structure reconstruction from incomplete coarse-grained to all atomistic models. Typical reconstruction schemes can be divided into four major steps. Usually, the first step is reconstruction of the protein backbone chain starting from the C-alpha trace. This is followed by side-chains rebuilding based on protein backbone geometry. Subsequently, hydrogen atoms can be reconstructed. Finally, the resulting all-atom models may require structure optimization. Many methods are available to perform each of these tasks. We discuss the available tools and their potential applications in integrative modeling pipelines that can transfer coarse-grained information from computational predictions, or experiment, to all atomistic structures.

In Situ Structural Restraints from Cross-Linking Mass Spectrometry in Human Mitochondria

The field of structural biology is increasingly focusing on studying proteins in situ, i.e., in their greater biological context. Cross-linking mass spectrometry (CLMS) is contributing to this effort, typically through the use of mass spectrometry (MS)-cleavable cross-linkers. Here, we apply the popular noncleavable cross-linker disuccinimidyl suberate (DSS) to human mitochondria and identify 5518 distance restraints between protein residues. Each distance restraint on proteins or their interactions provides structural information within mitochondria. Comparing these restraints to protein data bank (PDB)-deposited structures and comparative models reveals novel protein conformations. Our data suggest, among others, substrates and protein flexibility of mitochondrial heat shock proteins. Through this study, we bring forward two central points for the progression of CLMS towards large-scale in situ structural biology: First, clustered conflicts of cross-link data reveal in situ protein conformation states in contrast to error-rich individual conflicts. Second, noncleavable cross-linkers are compatible with proteome-wide studies.

Deciphering protein evolution and fitness landscapes with latent space models

Protein sequences contain rich information about protein evolution, fitness landscapes, and stability. Here we investigate how latent space models trained using variational auto-encoders can infer these properties from sequences. Using both simulated and real sequences, we show that the low dimensional latent space representation of sequences, calculated using the encoder model, captures both evolutionary and ancestral relationships between sequences. Together with experimental fitness data and Gaussian process regression, the latent space representation also enables learning the protein fitness landscape in a continuous low dimensional space. Moreover, the model is also useful in predicting protein mutational stability landscapes and quantifying the importance of stability in shaping protein evolution. Overall, we illustrate that the latent space models learned using variational auto-encoders provide a mechanism for exploration of the rich data contained in protein sequences regarding evolution, fitness and stability and hence are well-suited to help guide protein engineering efforts.

Computing the Pathogenicity of Wilson's Disease ATP7B Mutations: Implications for Disease Prevalence

Genetic variations in the gene encoding the copper-transport protein ATP7B are the primary cause of Wilson's disease. Controversially, clinical prevalence seems much smaller than the prevalence estimated by genetic screening tools, causing fear that many people are undiagnosed, although early diagnosis and treatment is essential. To address this issue, we benchmarked 16 state-of-the-art computational disease-prediction methods against established data of missense ATP7B mutations. Our results show that the quality of the methods varies widely. We show the importance of optimizing the threshold of the methods used to distinguish pathogenic from nonpathogenic mutations against data of clinically confirmed pathogenic and nonpathogenic mutations. We find that most methods use thresholds that predict too many ATP7B mutations to be pathogenic. Thus, our findings explain the current controversy on Wilson's disease prevalence because meta-analysis and text search methods include many computational estimates that lead to higher disease prevalence than clinically observed. As proteins and diseases differ widely, a one-size-fits-all threshold cannot distinguish pathogenic and nonpathogenic mutations efficiently, as shown here. We also show that amino acid changes with small evolutionary substitution probability, mainly due to amino acid volume, are more associated with the disease, implying a pathological effect on the conformational state of the protein, which could affect copper transport or adenosine triphosphate recognition and hydrolysis. These findings may be a first step toward a more quantitative genotype-phenotype relationship of Wilson's disease.

Deep-learning contact-map guided protein structure prediction in CASP13

We report the results of two fully automated structure prediction pipelines, "Zhang-Server" and "QUARK", in CASP13. The pipelines were built upon the C-I-TASSER and C-QUARK programs, which in turn are based on I-TASSER and QUARK but with three new modules: (a) a novel multiple sequence alignment (MSA) generation protocol to construct deep sequence-profiles for contact prediction; (b) an improved meta-method, NeBcon, which combines multiple contact predictors, including ResPRE that predicts contact-maps by coupling precision-matrices with deep residual convolutional neural-networks; and (c) an optimized contact potential to guide structure assembly simulations. For 50 CASP13 FM domains that lacked homologous templates, average TM-scores of the first models produced by C-I-TASSER and C-QUARK were 28% and 56% higher than those constructed by I-TASSER and QUARK, respectively. For the first time, contact-map predictions demonstrated usefulness on TBM domains with close homologous templates, where TM-scores of C-I-TASSER models were significantly higher than those of I-TASSER models with a P-value <.05. Detailed data analyses showed that the success of C-I-TASSER and C-QUARK was mainly due to the increased accuracy of deep-learning-based contact-maps, as well as the careful balance between sequence-based contact restraints, threading templates, and generic knowledge-based potentials. Nevertheless, challenges still remain for predicting quaternary structure of multi-domain proteins, due to the difficulties in domain partitioning and domain reassembly. In addition, contact prediction in terminal regions was often unsatisfactory due to the sparsity of MSAs. Development of new contact-based domain partitioning and assembly methods and training contact models on sparse MSAs may help address these issues.

Assessing the accuracy of contact predictions in CASP13

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

Ensembling multiple raw coevolutionary features with deep residual neural networks for contact-map prediction in CASP13

We report the results of residue-residue contact prediction of a new pipeline built purely on the learning of coevolutionary features in the CASP13 experiment. For a query sequence, the pipeline starts with the collection of multiple sequence alignments (MSAs) from multiple genome and metagenome sequence databases using two complementary Hidden Markov Model (HMM)-based searching tools. Three profile matrices, built on covariance, precision, and pseudolikelihood maximization respectively, are then created from the MSAs, which are used as the input features of a deep residual convolutional neural network architecture for contact-map training and prediction. Two ensembling strategies have been proposed to integrate the matrix features through end-to-end training and stacking, resulting in two complementary programs called TripletRes and ResTriplet, respectively. For the 31 free-modeling domains that do not have homologous templates in the PDB, TripletRes and ResTriplet generated comparable results with an average accuracy of 0.640 and 0.646, respectively, for the top L/5 long-range predictions, where 71% and 74% of the cases have an accuracy above 0.5. Detailed data analyses showed that the strength of the pipeline is due to the sensitive MSA construction and the advanced strategies for coevolutionary feature ensembling. Domain splitting was also found to help enhance the contact prediction performance. Nevertheless, contact models for tail regions, which often involve a high number of alignment gaps, and for targets with few homologous sequences are still suboptimal. Development of new approaches where the model is specifically trained on these regions and targets might help address these problems.

Coevolutive, evolutive and stochastic information in protein-protein interactions

Here, we investigate the contributions of coevolutive, evolutive and stochastic information in determining protein-protein interactions (PPIs) based on primary sequences of two interacting protein families A and B. Specifically, under the assumption that coevolutive information is imprinted on the interacting amino acids of two proteins in contrast to other (evolutive and stochastic) sources spread over their sequences, we dissect those contributions in terms of compensatory mutations at physically-coupled and uncoupled amino acids of A and B. We find that physically-coupled amino-acids at short range distances store the largest per-contact mutual information content, with a significant fraction of that content resulting from coevolutive sources alone. The information stored in coupled amino acids is shown further to discriminate multi-sequence alignments (MSAs) with the largest expectation fraction of PPI matches – a conclusion that holds against various definitions of intermolecular contacts and binding modes. When compared to the informational content resulting from evolution at long-range interactions, the mutual information in physically-coupled amino-acids is the strongest signal to distinguish PPIs derived from cospeciation and likely, the unique indication in case of molecular coevolution in independent genomes as the evolutive information must vanish for uncorrelated proteins.

Evolutionary coupling saturation mutagenesis: Coevolution-guided identification of distant sites influencing Bacillus naganoensis pullulanase activity

Pullulanases are well-known debranching enzymes hydrolyzing α-1,6-glycosidic linkages. To date, engineering of pullulanase is mainly focused on catalytic pocket or domain tailoring based on structure/sequence information. Saturation mutagenesis-involved directed evolution is, however, limited by the low number of mutational sites compatible with combinatorial libraries of feasible size. Using Bacillus naganoensis pullulanase as a target protein, here we introduce the 'evolutionary coupling saturation mutagenesis' (ECSM) approach: residue pair covariances are calculated to identify residues for saturation mutagenesis, focusing directed evolution on residue pairs playing important roles in natural evolution. Evolutionary coupling (EC) analysis identified seven residue pairs as evolutionary mutational hotspots. Subsequent saturation mutagenesis yielded variants with enhanced catalytic activity. The functional pairs apparently represent distant sites affecting enzyme activity.

PredMP: a web server for de novo prediction and visualization of membrane proteins

MOTIVATION

PredMP is the first web service, to our knowledge, that aims at de novo prediction of the membrane protein (MP) 3D structure followed by the embedding of the MP into the lipid bilayer for visualization. Our approach is based on a high-throughput Deep Transfer Learning (DTL) method that first predicts MP contacts by learning from non-MPs and then predicts the 3D model of the MP using the predicted contacts as distance restraints. This algorithm is derived from our previous Deep Learning (DL) method originally developed for soluble protein contact prediction, which has been officially ranked No. 1 in CASP12. The DTL framework in our approach overcomes the challenge that there are only a limited number of solved MP structures for training the deep learning model. There are three modules in the PredMP server: (i) The DTL framework followed by the contact-assisted folding protocol has already been implemented in RaptorX-Contact, which serves as the key module for 3D model generation; (ii) The 1D annotation module, implemented in RaptorX-Property, is used to predict the secondary structure and disordered regions; and (iii) the visualization module to display the predicted MPs embedded in the lipid bilayer guided by the predicted transmembrane topology.

RESULTS

Tested on 510 non-redundant MPs, our server predicts correct folds for ∼290 MPs, which significantly outperforms existing methods. Tested on a blind and live benchmark CAMEO from September 2016 to January 2018, PredMP can successfully model all 10 MPs belonging to the hard category.

AVAILABILITY AND IMPLEMENTATION

PredMP is freely accessed on the web at http://www.predmp.com.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

AlphaFold at CASP13

SUMMARY

Computational prediction of protein structure from sequence is broadly viewed as a foundational problem of biochemistry and one of the most difficult challenges in bioinformatics. Once every two years the Critical Assessment of protein Structure Prediction (CASP) experiments are held to assess the state of the art in the field in a blind fashion, by presenting predictor groups with protein sequences whose structures have been solved but have not yet been made publicly available. The first CASP was organized in 1994, and the latest, CASP13, took place last December, when for the first time the industrial laboratory DeepMind entered the competition. DeepMind's entry, AlphaFold, placed first in the Free Modeling (FM) category, which assesses methods on their ability to predict novel protein folds (the Zhang group placed first in the Template-Based Modeling (TBM) category, which assess methods on predicting proteins whose folds are related to ones already in the Protein Data Bank.) DeepMind's success generated significant public interest. Their approach builds on two ideas developed in the academic community during the preceding decade: (i) the use of co-evolutionary analysis to map residue co-variation in protein sequence to physical contact in protein structure, and (ii) the application of deep neural networks to robustly identify patterns in protein sequence and co-evolutionary couplings and convert them into contact maps. In this Letter, we contextualize the significance of DeepMind's entry within the broader history of CASP, relate AlphaFold's methodological advances to prior work, and speculate on the future of this important problem.

Fueling ab initio folding with marine metagenomics enables structure and function predictions of new protein families

INTRODUCTION

The ocean microbiome represents one of the largest microbiomes and produces nearly half of the primary energy on the planet through photosynthesis or chemosynthesis. Using recent advances in marine genomics, we explore new applications of oceanic metagenomes for protein structure and function prediction.

RESULTS

By processing 1.3 TB of high-quality reads from the Tara Oceans data, we obtain 97 million non-redundant genes. Of the 5721 Pfam families that lack experimental structures, 2801 have at least one member associated with the oceanic metagenomics dataset. We apply C-QUARK, a deep-learning contact-guided ab initio structure prediction pipeline, to model 27 families, where 20 are predicted to have a reliable fold with estimated template modeling score (TM-score) at least 0.5. Detailed analyses reveal that the abundance of microbial genera in the ocean is highly correlated to the frequency of occurrence in the modeled Pfam families, suggesting the significant role of the Tara Oceans genomes in the contact-map prediction and subsequent ab initio folding simulations. Of interesting note, PF15461, which has a majority of members coming from ocean-related bacteria, is identified as an important photosynthetic protein by structure-based function annotations. The pipeline is extended to a set of 417 Pfam families, built on the combination of Tara with other metagenomics datasets, which results in 235 families with an estimated TM-score over 0.5.

CONCLUSIONS

These results demonstrate a new avenue to improve the capacity of protein structure and function modeling through marine metagenomics, especially for difficult proteins with few homologous sequences.

Structural diversity in the Mycobacteria DUF3349 superfamily

A protein superfamily with a "Domain of Unknown Function,", DUF3349 (PF11829), is present predominately in Mycobacterium and Rhodococcus bacterial species suggesting that these proteins may have a biological function unique to these bacteria. We previously reported the inaugural structure of a DUF3349 superfamily member, Mycobacterium tuberculosis Rv0543c. Here, we report the structures determined for three additional DUF3349 proteins: Mycobacterium smegmatis MSMEG_1063 and MSMEG_1066 and Mycobacterium abscessus MAB_3403c. Like Rv0543c, the NMR solution structure of MSMEG_1063 revealed a monomeric five α-helix bundle with a similar overall topology. Conversely, the crystal structure of MSMEG_1066 revealed a five α-helix protein with a strikingly different topology and a tetrameric quaternary structure that was confirmed by size exclusion chromatography. The NMR solution structure of a fourth member of the DUF3349 superfamily, MAB_3403c, with 18 residues missing at the N-terminus, revealed a monomeric α-helical protein with a folding topology similar to the three C-terminal helices in the protomer of the MSMEG_1066 tetramer. These structures, together with a GREMLIN-based bioinformatics analysis of the DUF3349 primary amino acid sequences, suggest two subfamilies within the DUF3349 family. The division of the DUF3349 into two distinct subfamilies would have been lost if structure solution had stopped with the first structure in the DUF3349 family, highlighting the insights generated by solving multiple structures within a protein superfamily. Future studies will determine if the structural diversity at the tertiary and quaternary levels in the DUF3349 protein superfamily have functional roles in Mycobacteria and Rhodococcus species with potential implications for structure-based drug discovery.

Predicting protein inter-residue contacts using composite likelihood maximization and deep learning

BACKGROUND

Accurate prediction of inter-residue contacts of a protein is important to calculating its tertiary structure. Analysis of co-evolutionary events among residues has been proved effective in inferring inter-residue contacts. The Markov random field (MRF) technique, although being widely used for contact prediction, suffers from the following dilemma: the actual likelihood function of MRF is accurate but time-consuming to calculate; in contrast, approximations to the actual likelihood, say pseudo-likelihood, are efficient to calculate but inaccurate. Thus, how to achieve both accuracy and efficiency simultaneously remains a challenge.

RESULTS

In this study, we present such an approach (called clmDCA) for contact prediction. Unlike plmDCA using pseudo-likelihood, i.e., the product of conditional probability of individual residues, our approach uses composite-likelihood, i.e., the product of conditional probability of all residue pairs. Composite likelihood has been theoretically proved as a better approximation to the actual likelihood function than pseudo-likelihood. Meanwhile, composite likelihood is still efficient to maximize, thus ensuring the efficiency of clmDCA. We present comprehensive experiments on popular benchmark datasets, including PSICOV dataset and CASP-11 dataset, to show that: i) clmDCA alone outperforms the existing MRF-based approaches in prediction accuracy. ii) When equipped with deep learning technique for refinement, the prediction accuracy of clmDCA was further significantly improved, suggesting the suitability of clmDCA for subsequent refinement procedure. We further present a successful application of the predicted contacts to accurately build tertiary structures for proteins in the PSICOV dataset.

CONCLUSIONS

Composite likelihood maximization algorithm can efficiently estimate the parameters of Markov Random Fields and can improve the prediction accuracy of protein inter-residue contacts.

Accurate prediction of protein contact maps by coupling residual two-dimensional bidirectional long short-term memory with convolutional neural networks

Motivation

Accurate prediction of a protein contact map depends greatly on capturing as much contextual information as possible from surrounding residues for a target residue pair. Recently, ultra-deep residual convolutional networks were found to be state-of-the-art in the latest Critical Assessment of Structure Prediction techniques (CASP12) for protein contact map prediction by attempting to provide a protein-wide context at each residue pair. Recurrent neural networks have seen great success in recent protein residue classification problems due to their ability to propagate information through long protein sequences, especially Long Short-Term Memory (LSTM) cells. Here, we propose a novel protein contact map prediction method by stacking residual convolutional networks with two-dimensional residual bidirectional recurrent LSTM networks, and using both one-dimensional sequence-based and two-dimensional evolutionary coupling-based information.

Results

We show that the proposed method achieves a robust performance over validation and independent test sets with the Area Under the receiver operating characteristic Curve (AUC) > 0.95 in all tests. When compared to several state-of-the-art methods for independent testing of 228 proteins, the method yields an AUC value of 0.958, whereas the next-best method obtains an AUC of 0.909. More importantly, the improvement is over contacts at all sequence-position separations. Specifically, a 8.95%, 5.65% and 2.84% increase in precision were observed for the top L∕10 predictions over the next best for short, medium and long-range contacts, respectively. This confirms the usefulness of ResNets to congregate the short-range relations and 2D-BRLSTM to propagate the long-range dependencies throughout the entire protein contact map 'image'.

Availability and implementation

SPOT-Contact server url: http://sparks-lab.org/jack/server/SPOT-Contact/.

Supplementary information

Supplementary data are available at Bioinformatics online.

Detecting distant-homology protein structures by aligning deep neural-network based contact maps

Accurate prediction of atomic-level protein structure is important for annotating the biological functions of protein molecules and for designing new compounds to regulate the functions. Template-based modeling (TBM), which aims to construct structural models by copying and refining the structural frameworks of other known proteins, remains the most accurate method for protein structure prediction. Due to the difficulty in recognizing distant-homology templates, however, the accuracy of TBM decreases rapidly when the evolutionary relationship between the query and template vanishes. In this study, we propose a new method, CEthreader, which first predicts residue-residue contacts by coupling evolutionary precision matrices with deep residual convolutional neural-networks. The predicted contact maps are then integrated with sequence profile alignments to recognize structural templates from the PDB. The method was tested on two independent benchmark sets consisting collectively of 1,153 non-homologous protein targets, where CEthreader detected 176% or 36% more correct templates with a TM-score >0.5 than the best state-of-the-art profile- or contact-based threading methods, respectively, for the Hard targets that lacked homologous templates. Moreover, CEthreader was able to identify 114% or 20% more correct templates with the same Fold as the query, after excluding structures from the same SCOPe Superfamily, than the best profile- or contact-based threading methods. Detailed analyses show that the major advantage of CEthreader lies in the efficient coupling of contact maps with profile alignments, which helps recognize global fold of protein structures when the homologous relationship between the query and template is weak. These results demonstrate an efficient new strategy to combine ab initio contact map prediction with profile alignments to significantly improve the accuracy of template-based structure prediction, especially for distant-homology proteins.

Despite decades of effort in computational method development, template-based modeling (TBM) still remains the most reliable approach to high-resolution protein structure prediction. Previous studies have shown that the PDB library is complete for single-domain proteins and TBM is in principle sufficient to solve the structure prediction problem if the most similar structure in the PDB could be reliably identified and used as template for model reconstruction. But in reality, the success of TBM depends on the availability of closely-homologous templates, where its accuracy and reliability decrease sharply when the evolutionary relationship between query and template becomes more distant. We developed a new threading approach, CEthreader, which allows for dynamic programing alignments of predicted contact-maps through eigen-decomposition. The large-scale benchmark tests show that the coupling of contact map with profile and secondary structure alignments through the proposed protocol can significantly improve the accuracy of template recognition for distantly-homologous protein targets.

Recent developments in deep learning applied to protein structure prediction

Although many structural bioinformatics tools have been using neural network models for a long time, deep neural network (DNN) models have attracted considerable interest in recent years. Methods employing DNNs have had a significant impact in recent CASP experiments, notably in CASP12 and especially CASP13. In this article, we offer a brief introduction to some of the key principles and properties of DNN models and discuss why they are naturally suited to certain problems in structural bioinformatics. We also briefly discuss methodological improvements that have enabled these successes. Using the contact prediction task as an example, we also speculate why DNN models are able to produce reasonably accurate predictions even in the absence of many homologues for a given target sequence, a result that can at first glance appear surprising given the lack of input information. We end on some thoughts about how and why these types of models can be so effective, as well as a discussion on potential pitfalls.

Remote homology searches identify bacterial homologues of eukaryotic lipid transfer proteins, including Chorein-N domains in TamB and AsmA and Mdm31p

Background

All cells rely on lipids for key functions. Lipid transfer proteins allow lipids to exit the hydrophobic environment of bilayers, and cross aqueous spaces. One lipid transfer domain fold present in almost all eukaryotes is the TUbular LIPid binding (TULIP) domain. Three TULIP families have been identified in bacteria (P47, OrfX2 and YceB), but their homology to eukaryotic proteins is too low to specify a common origin. Another recently described eukaryotic lipid transfer domain in VPS13 and ATG2 is Chorein-N, which has no known bacterial homologues. There has been no systematic search for bacterial TULIPs or Chorein-N domains.

Results

Remote homology predictions for bacterial TULIP domains using HHsearch identified four new TULIP domains in three bacterial families. DUF4403 is a full length pseudo-dimeric TULIP with a 6 strand β-meander dimer interface like eukaryotic TULIPs. A similar sheet is also present in YceB, suggesting it homo-dimerizes. TULIP domains were also found in DUF2140 and in the C-terminus DUF2993. Remote homology predictions for bacterial Chorein-N domains identified strong hits in the N-termini of AsmA and TamB in diderm bacteria, which are related to Mdm31p in eukaryotic mitochondria. The N-terminus of DUF2993 has a Chorein-N domain adjacent to its TULIP domain.

Conclusions

TULIP lipid transfer domains are widespread in bacteria. Chorein-N domains are also found in bacteria, at the N-terminus of multiple proteins in the intermembrane space of diderms (AsmA, TamB and their relatives) and in Mdm31p, a protein that is likely to have evolved from an AsmA/TamB-like protein in the endosymbiotic mitochondrial ancestor. This indicates that both TULIP and Chorein-N lipid transfer domains may have originated in bacteria.

Phylogenetic weighting does little to improve the accuracy of evolutionary coupling analyses

Homologous sequence alignments contain important information about the constraints that shape protein family evolution. Correlated changes between different residues, for instance, can be highly predictive of physical contacts within three-dimensional structures. Detecting such co-evolutionary signals via direct coupling analysis is particularly challenging given the shared phylogenetic history and uneven sampling of different lineages from which protein sequences are derived. Current best practices for mitigating such effects include sequence-identity-based weighting of input sequences and post-hoc re-scaling of evolutionary coupling scores. However, numerous weighting schemes have been previously developed for other applications, and it is unknown whether any of these schemes may better account for phylogenetic artifacts in evolutionary coupling analyses. Here, we show across a dataset of 150 diverse protein families that the current best practices out-perform several alternative sequence- and tree-based weighting methods. Nevertheless, we find that sequence weighting in general provides only a minor benefit relative to post-hoc transformations that re-scale the derived evolutionary couplings. While our findings do not rule out the possibility that an as-yet-untested weighting method may show improved results, the similar predictive accuracies that we observe across distinct weighting methods suggests that there may be little room for further improvement on top of existing strategies.

A multi-scale coevolutionary approach to predict interactions between protein domains

Interacting proteins and protein domains coevolve on multiple scales, from their correlated presence across species, to correlations in amino-acid usage. Genomic databases provide rapidly growing data for variability in genomic protein content and in protein sequences, calling for computational predictions of unknown interactions. We first introduce the concept of direct phyletic couplings, based on global statistical models of phylogenetic profiles. They strongly increase the accuracy of predicting pairs of related protein domains beyond simpler correlation-based approaches like phylogenetic profiling (80% vs. 30-50% positives out of the 1000 highest-scoring pairs). Combined with the direct coupling analysis of inter-protein residue-residue coevolution, we provide multi-scale evidence for direct but unknown interaction between protein families. An in-depth discussion shows these to be biologically sensible and directly experimentally testable. Negative phyletic couplings highlight alternative solutions for the same functionality, including documented cases of convergent evolution. Thereby our work proves the strong potential of global statistical modeling approaches to genome-wide coevolutionary analysis, far beyond the established use for individual protein complexes and domain-domain interactions.

Template-based modeling by ClusPro in CASP13 and the potential for using co-evolutionary information in docking

As a participant in the joint CASP13-CAPRI46 assessment, the ClusPro server debuted its new template-based modeling functionality. The addition of this feature, called ClusPro TBM, was motivated by the previous CASP-CAPRI assessments and by the proven ability of template-based methods to produce higher-quality models, provided templates are available. In prior assessments, ClusPro submissions consisted of models that were produced via free docking of pre-generated homology models. This method was successful in terms of the number of acceptable predictions across targets; however, analysis of results showed that purely template-based methods produced a substantially higher number of medium-quality models for targets for which there were good templates available. The addition of template-based modeling has expanded ClusPro's ability to produce higher accuracy predictions, primarily for homomeric but also for some heteromeric targets. Here we review the newest additions to the ClusPro web server and discuss examples of CASP-CAPRI targets that continue to drive further development. We also describe ongoing work not yet implemented in the server. This includes the development of methods to improve template-based models and the use of co-evolutionary information for data-assisted free docking.

An Improved Method for Modeling Voltage-Gated Ion Channels at Atomic Accuracy Applied to Human Cav Channels

Voltage-gated ion channels (VGICs) are associated with hundreds of human diseases. To date, 3D structural models of human VGICs have not been reported. We developed a 3D structural integrity metric to rank the accuracy of all VGIC structures deposited in the PDB. The metric revealed inaccuracies in structural models built from recent single-particle, non-crystalline cryo-electron microscopy maps and enabled the building of highly accurate homology models of human Ca_v channel α₁ subunits at atomic resolution. Human Ca_v Mendelian mutations mostly located to segments involved in the mechanism of voltage sensing and gating within the 3D structure, with multiple mutations targeting equivalent 3D structural locations despite eliciting distinct clinical phenotypes. The models also revealed that the architecture of the ion selectivity filter is highly conserved from bacteria to humans and between sodium and calcium VGICs.

Targeted isolation and cultivation of uncultivated bacteria by reverse genomics

Most microorganisms from all taxonomic levels are uncultured. Single-cell genomes and metagenomes continue to increase the known diversity of Bacteria and Archaea, but while ‘omics can be used to infer physiological or ecological roles for species in a community, most of those hypothetical roles remain unvalidated. Here we report an approach to capture specific microorganisms from complex communities into pure cultures using genome-informed antibody engineering. We apply our reverse genomics approach to isolate and sequence single cells and to cultivate three different species-level lineages of human oral Saccharibacteria/TM7. Using our pure cultures we show that all three saccharibacteria species are epibionts of diverse Actinobacteria. We also isolate and cultivate human oral SR1 bacteria, which are members of a lineage of previously uncultured bacteria. Reverse-genomics-enabled cultivation of microorganisms can be applied to any species from any environment and has the potential to unlock the isolation, cultivation and characterization of species from as-yet-uncultured branches of the microbial tree of life.

Tissue-Specific Subcellular Localization Prediction Using Multi-Label Markov Random Fields

The understanding of subcellular localization (SCL) of proteins and proteome variation in the different tissues and organs of the human body are two crucial aspects for increasing our knowledge of the dynamic rules of proteins, the cell biology, and the mechanism of diseases. Although there have been tremendous contributions to these two fields independently, the lack of knowledge of the variation of spatial distribution of proteins in the different tissues still exists. Here, we proposed an approach that allows predicting protein SCL on tissue specificity through the use of tissue-specific functional associations and physical protein-protein interactions (PPIs). We applied our previously developed Bayesian collective Markov random fields (BCMRFs) on tissue-specific protein-protein interaction network (PPI network) for nine types of tissues focusing on eight high-level SCL. The evaluated results demonstrate the strength of our approach in predicting tissue-specific SCL. We identified 1,314 proteins that their SCL were previously proven cell line dependent. We predicted 549 novel tissue-specific localized candidate proteins while some of them were validated via text-mining.

Accurate Classification of Biological and non-Biological Interfaces in Protein Crystal Structures using Subtle Covariation Signals

Proteins often work as oligomers or multimers in vivo. Therefore, elucidating their oligomeric or multimeric form (quaternary structure) is crucially important to ascertain their function. X-ray crystal structures of numerous proteins have been accumulated, providing information related to their biological units. Extracting information of biological units from protein crystal structures represents a meaningful task for modern biology. Nevertheless, although many methods have been proposed for identifying biological units appearing in protein crystal structures, it is difficult to distinguish biological protein-protein interfaces from crystallographic ones. Therefore, our simple but highly accurate classifier was developed to infer biological units in protein crystal structures using large amounts of protein sequence information and a modern contact prediction method to exploit covariation signals (CSs) in proteins. We demonstrate that our proposed method is promising even for weak signals of biological interfaces. We also discuss the relation between classification accuracy and conservation of biological units, and illustrate how the selection of sequences included in multiple sequence alignments as sources for obtaining CSs affects the results. With increased amounts of sequence data, the proposed method is expected to become increasingly useful.

Analysis of several key factors influencing deep learning-based inter-residue contact prediction

MOTIVATION

Deep learning has become the dominant technology for protein contact prediction. However, the factors that affect the performance of deep learning in contact prediction have not been systematically investigated.

RESULTS

We analyzed the results of our three deep learning-based contact prediction methods (MULTICOM-CLUSTER, MULTICOM-CONSTRUCT and MULTICOM-NOVEL) in the CASP13 experiment and identified several key factors [i.e. deep learning technique, multiple sequence alignment (MSA), distance distribution prediction and domain-based contact integration] that influenced the contact prediction accuracy. We compared our convolutional neural network (CNN)-based contact prediction methods with three coevolution-based methods on 75 CASP13 targets consisting of 108 domains. We demonstrated that the CNN-based multi-distance approach was able to leverage global coevolutionary coupling patterns comprised of multiple correlated contacts for more accurate contact prediction than the local coevolution-based methods, leading to a substantial increase of precision by 19.2 percentage points. We also tested different alignment methods and domain-based contact prediction with the deep learning contact predictors. The comparison of the three methods showed deeper sequence alignments and the integration of domain-based contact prediction with the full-length contact prediction improved the performance of contact prediction. Moreover, we demonstrated that the domain-based contact prediction based on a novel ab initio approach of parsing domains from MSAs alone without using known protein structures was a simple, fast approach to improve contact prediction. Finally, we showed that predicting the distribution of inter-residue distances in multiple distance intervals could capture more structural information and improve binary contact prediction.

AVAILABILITY AND IMPLEMENTATION

https://github.com/multicom-toolbox/DNCON2/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Distance-based protein folding powered by deep learning

Direct coupling analysis (DCA) for protein folding has made very good progress, but it is not effective for proteins that lack many sequence homologs, even coupled with time-consuming conformation sampling with fragments. We show that we can accurately predict interresidue distance distribution of a protein by deep learning, even for proteins with ∼60 sequence homologs. Using only the geometric constraints given by the resulting distance matrix we may construct 3D models without involving extensive conformation sampling. Our method successfully folded 21 of the 37 CASP12 hard targets with a median family size of 58 effective sequence homologs within 4 h on a Linux computer of 20 central processing units. In contrast, DCA-predicted contacts cannot be used to fold any of these hard targets in the absence of extensive conformation sampling, and the best CASP12 group folded only 11 of them by integrating DCA-predicted contacts into fragment-based conformation sampling. Rigorous experimental validation in CASP13 shows that our distance-based folding server successfully folded 17 of 32 hard targets (with a median family size of 36 sequence homologs) and obtained 70% precision on the top L/5 long-range predicted contacts. The latest experimental validation in CAMEO shows that our server predicted correct folds for 2 membrane proteins while all of the other servers failed. These results demonstrate that it is now feasible to predict correct fold for many more proteins lack of similar structures in the Protein Data Bank even on a personal computer.

NMR-assisted protein structure prediction with MELDxMD

We describe the performance of MELD-accelerated molecular dynamics (MELDxMD) in determining protein structures in the NMR-data-assisted category in CASP13. Seeded from web server predictions, MELDxMD was found best in the NMR category, over 17 targets, outperforming the next-best groups by a factor of ~4 in z-score. MELDxMD gives ensembles, not single structures; succeeds on a 326-mer, near the current upper limit for NMR structures; and predicts structures that match experimental residual dipolar couplings even though the only NMR-derived data used in the simulations was NOE-based ambiguous atom-atom contacts and backbone dihedrals. MELD can use noisy and ambiguous experimental information to reduce the MD search space. We believe MELDxMD is a promising method for determining protein structures from NMR data.

ComplexContact: a web server for inter-protein contact prediction using deep learning

ComplexContact (http://raptorx2.uchicago.edu/ComplexContact/) is a web server for sequence-based interfacial residue-residue contact prediction of a putative protein complex. Interfacial residue-residue contacts are critical for understanding how proteins form complex and interact at residue level. When receiving a pair of protein sequences, ComplexContact first searches for their sequence homologs and builds two paired multiple sequence alignments (MSA), then it applies co-evolution analysis and a CASP-winning deep learning (DL) method to predict interfacial contacts from paired MSAs and visualizes the prediction as an image. The DL method was originally developed for intra-protein contact prediction and performed the best in CASP12. Our large-scale experimental test further shows that ComplexContact greatly outperforms pure co-evolution methods for inter-protein contact prediction, regardless of the species.

High-accuracy refinement using Rosetta in CASP13

Because proteins generally fold to their lowest free energy states, energy-guided refinement in principle should be able to systematically improve the quality of protein structure models generated using homologous structure or co-evolution derived information. However, because of the high dimensionality of the search space, there are far more ways to degrade the quality of a near native model than to improve it, and hence, refinement methods are very sensitive to energy function errors. In the 13th Critial Assessment of techniques for protein Structure Prediction (CASP13), we sought to carry out a thorough search for low energy states in the neighborhood of a starting model using restraints to avoid straying too far. The approach was reasonably successful in improving both regions largely incorrect in the starting models as well as core regions that started out closer to the correct structure. Models with GDT-HA over 70 were obtained for five targets and for one of those, an accuracy of 0.5 å backbone root-mean-square deviation (RMSD) was achieved. An important current challenge is to improve performance in refining oligomers and larger proteins, for which the search problem remains extremely difficult.