PSICOV: precise structural contact prediction using sparse inverse covariance estimation on large multiple sequence alignments

Modeling sequence-space exploration and emergence of epistatic signals in protein evolution

During their evolution, proteins explore sequence space via an interplay between random mutations and phenotypic selection. Here, we build upon recent progress in reconstructing data-driven fitness landscapes for families of homologous proteins, to propose stochastic models of experimental protein evolution. These models predict quantitatively important features of experimentally evolved sequence libraries, like fitness distributions and position-specific mutational spectra. They also allow us to efficiently simulate sequence libraries for a vast array of combinations of experimental parameters like sequence divergence, selection strength, and library size. We showcase the potential of the approach in reanalyzing two recent experiments to determine protein structure from signals of epistasis emerging in experimental sequence libraries. To be detectable, these signals require sufficiently large and sufficiently diverged libraries. Our modeling framework offers a quantitative explanation for different outcomes of recently published experiments. Furthermore, we can forecast the outcome of time- and resource-intensive evolution experiments, opening thereby a way to computationally optimize experimental protocols.

Intelligent host engineering for metabolic flux optimisation in biotechnology

Optimising the function of a protein of length N amino acids by directed evolution involves navigating a 'search space' of possible sequences of some 20N. Optimising the expression levels of P proteins that materially affect host performance, each of which might also take 20 (logarithmically spaced) values, implies a similar search space of 20P. In this combinatorial sense, then, the problems of directed protein evolution and of host engineering are broadly equivalent. In practice, however, they have different means for avoiding the inevitable difficulties of implementation. The spare capacity exhibited in metabolic networks implies that host engineering may admit substantial increases in flux to targets of interest. Thus, we rehearse the relevant issues for those wishing to understand and exploit those modern genome-wide host engineering tools and thinking that have been designed and developed to optimise fluxes towards desirable products in biotechnological processes, with a focus on microbial systems. The aim throughput is 'making such biology predictable'. Strategies have been aimed at both transcription and translation, especially for regulatory processes that can affect multiple targets. However, because there is a limit on how much protein a cell can produce, increasing kcat in selected targets may be a better strategy than increasing protein expression levels for optimal host engineering.

Extending the Horizon of Homology Detection with Coevolution-based Structure Prediction

Traditional sequence analysis algorithms fail to identify distant homologies when they lie beyond a detection horizon. In this review, we discuss how co-evolution-based contact and distance prediction methods are pushing back this homology detection horizon, thereby yielding new functional insights and experimentally testable hypotheses. Based on correlated substitutions, these methods divine three-dimensional constraints among amino acids in protein sequences that were previously devoid of all annotated domains and repeats. The new algorithms discern hidden structure in an otherwise featureless sequence landscape. Their revelatory impact promises to be as profound as the use, by archaeologists, of ground-penetrating radar to discern long-hidden, subterranean structures. As examples of this, we describe how triplicated structures reflecting longin domains in MON1A-like proteins, or UVR-like repeats in DISC1, emerge from their predicted contact and distance maps. These methods also help to resolve structures that do not conform to a "beads-on-a-string" model of protein domains. In one such example, we describe CFAP298 whose ubiquitin-like domain was previously challenging to perceive owing to a large sequence insertion within it. More generally, the new algorithms permit an easier appreciation of domain families and folds whose evolution involved structural insertion or rearrangement. As we exemplify with α1-antitrypsin, coevolution-based predicted contacts may also yield insights into protein dynamics and conformational change. This new combination of structure prediction (using innovative co-evolution based methods) and homology inference (using more traditional sequence analysis approaches) shows great promise for bringing into view a sea of evolutionary relationships that had hitherto lain far beyond the horizon of homology detection.

Sparse generative modeling via parameter reduction of Boltzmann machines: Application to protein-sequence families

Boltzmann machines (BMs) are widely used as generative models. For example, pairwise Potts models (PMs), which are instances of the BM class, provide accurate statistical models of families of evolutionarily related protein sequences. Their parameters are the local fields, which describe site-specific patterns of amino acid conservation, and the two-site couplings, which mirror the coevolution between pairs of sites. This coevolution reflects structural and functional constraints acting on protein sequences during evolution. The most conservative choice to describe the coevolution signal is to include all possible two-site couplings into the PM. This choice, typical of what is known as Direct Coupling Analysis, has been successful for predicting residue contacts in the three-dimensional structure, mutational effects, and generating new functional sequences. However, the resulting PM suffers from important overfitting effects: many couplings are small, noisy, and hardly interpretable; the PM is close to a critical point, meaning that it is highly sensitive to small parameter perturbations. In this work, we introduce a general parameter-reduction procedure for BMs, via a controlled iterative decimation of the less statistically significant couplings, identified by an information-based criterion that selects either weak or statistically unsupported couplings. For several protein families, our procedure allows one to remove more than 90% of the PM couplings, while preserving the predictive and generative properties of the original dense PM, and the resulting model is far away from criticality, hence more robust to noise.

Protein sequence-to-structure learning: Is this the end(-to-end revolution)?

The potential of deep learning has been recognized in the protein structure prediction community for some time, and became indisputable after CASP13. In CASP14, deep learning has boosted the field to unanticipated levels reaching near-experimental accuracy. This success comes from advances transferred from other machine learning areas, as well as methods specifically designed to deal with protein sequences and structures, and their abstractions. Novel emerging approaches include (i) geometric learning, that is, learning on representations such as graphs, three-dimensional (3D) Voronoi tessellations, and point clouds; (ii) pretrained protein language models leveraging attention; (iii) equivariant architectures preserving the symmetry of 3D space; (iv) use of large meta-genome databases; (v) combinations of protein representations; and (vi) finally truly end-to-end architectures, that is, differentiable models starting from a sequence and returning a 3D structure. Here, we provide an overview and our opinion of the novel deep learning approaches developed in the last 2 years and widely used in CASP14.

When homologous sequences meet structural decoys: Accurate contact prediction by tFold in CASP14-(tFold for CASP14 contact prediction)

In this paper, we report our tFold framework's performance on the inter-residue contact prediction task in the 14th Critical Assessment of protein Structure Prediction (CASP14). Our tFold framework seamlessly combines both homologous sequences and structural decoys under an ultra-deep network architecture. Squeeze-excitation and axial attention mechanisms are employed to effectively capture inter-residue interactions. In CASP14, our best predictor achieves 41.78% in the averaged top-L precision for long-range contacts for all the 22 free-modeling (FM) targets, and ranked 1st among all the 60 participating teams. The tFold web server is now freely available at: https://drug.ai.tencent.com/console/en/tfold.

Accurate prediction of inter-protein residue-residue contacts for homo-oligomeric protein complexes

Protein-protein interactions play a fundamental role in all cellular processes. Therefore, determining the structure of protein-protein complexes is crucial to understand their molecular mechanisms and develop drugs targeting the protein-protein interactions. Recently, deep learning has led to a breakthrough in intra-protein contact prediction, achieving an unusual high accuracy in recent Critical Assessment of protein Structure Prediction (CASP) structure prediction challenges. However, due to the limited number of known homologous protein-protein interactions and the challenge to generate joint multiple sequence alignments of two interacting proteins, the advances in inter-protein contact prediction remain limited. Here, we have proposed a deep learning model to predict inter-protein residue-residue contacts across homo-oligomeric protein interfaces, named as DeepHomo. Unlike previous deep learning approaches, we integrated intra-protein distance map and inter-protein docking pattern, in addition to evolutionary coupling, sequence conservation, and physico-chemical information of monomers. DeepHomo was extensively tested on both experimentally determined structures and realistic CASP-Critical Assessment of Predicted Interaction (CAPRI) targets. It was shown that DeepHomo achieved a high precision of >60% for the top predicted contact and outperformed state-of-the-art direct-coupling analysis and machine learning-based approaches. Integrating predicted inter-chain contacts into protein-protein docking significantly improved the docking accuracy on the benchmark dataset of realistic homo-dimeric targets from CASP-CAPRI experiments. DeepHomo is available at http://huanglab.phys.hust.edu.cn/DeepHomo/.

Empirical Potential Energy Function Toward ab Initio Folding G Protein-Coupled Receptors

Approximately 40-50 percent of all drugs targets are G protein-coupled receptors (GPCRs). Three-dimensional structure of GPCRs is important to probe their biophysical and biochemical functions and their pharmaceutical applications. Lacking reliable and high quality free function is one of the ugent problems of computational predicting the three-dimensional structure in this community. We proposed a GPCR-specified energy function composed of four novel empirical potential energy terms: a two-dimensional contact energy force field, knowledge-based helix pair connection distance energy term, knowledge-based helix pair angle restraint energy term and a disulfide bond energy term. To validate the energy function, we employed an ab initio GPCR three-dimensional structure predictor to test if the energy function improved the accuracy of prediction. We evaluated 28 solved GPCRs and found that 21(75 percent) targets were correctly folded (TM-score>0.5). Also, the average TM-score using the energy function was 0.54, which was improved 134 percent than the TM-score 0.23 for MODELLER energy function and 170 percent than the TM-score 0.20 for Rosetta membrane energy function. The results confirmed that our empirical potential energy function toward ab initio folding is competitive to state-of-the-art solutions for structural prediction of GPCRs.

Improving fragment-based ab initio protein structure assembly using low-accuracy contact-map predictions

Sequence-based contact prediction has shown considerable promise in assisting non-homologous structure modeling, but it often requires many homologous sequences and a sufficient number of correct contacts to achieve correct folds. Here, we developed a method, C-QUARK, that integrates multiple deep-learning and coevolution-based contact-maps to guide the replica-exchange Monte Carlo fragment assembly simulations. The method was tested on 247 non-redundant proteins, where C-QUARK could fold 75% of the cases with TM-scores (template-modeling scores) ≥0.5, which was 2.6 times more than that achieved by QUARK. For the 59 cases that had either low contact accuracy or few homologous sequences, C-QUARK correctly folded 6 times more proteins than other contact-based folding methods. C-QUARK was also tested on 64 free-modeling targets from the 13th CASP (critical assessment of protein structure prediction) experiment and had an average GDT_TS (global distance test) score that was 5% higher than the best CASP predictors. These data demonstrate, in a robust manner, the progress in modeling non-homologous protein structures using low-accuracy and sparse contact-map predictions.

New approaches to predict the effect of co-occurring variants on protein characteristics

Predicting the effect of a mutated gene before the onset of symptoms of genetic diseases would greatly facilitate diagnosis and potentiate early intervention. There have been myriad attempts to predict the effects of single-nucleotide variants. However, the applicability of these efforts does not scale to co-occurring variants. Furthermore, an increasing number of protein therapeutics contain co-occurring nucleotide variations, adding uncertainty during development to the safety and efficiency of these drugs. Co-occurring nucleotide variants may often have synergistic, additive, or antagonistic effects on protein attributes, further complicating the task of outcome prediction. We tested four models based on the cooperative and antagonistic effects of co-occurring variants to predict pathogenicity and effectiveness of protein therapeutics. A total of 30 attributes, including amino acid and nucleotide features, as well as existing single-variant effect prediction tools, were considered on the basis of previous studies on single-nucleotide variants. Importantly, the effects of synonymous variants, often seen in protein therapeutics, were also included in our models. We used 12 datasets of people with monogenic diseases and controls with co-occurring genetic variants to evaluate the accuracy of our models, accomplishing a degree of accuracy comparable to that of prediction tools for single-nucleotide variants. More importantly, our framework is generalizable to new, well-curated datasets of monogenic diseases and new variant scoring tools. This approach successfully assists in addressing the challenging task of predicting the effect of co-occurring variants on pathogenicity and protein effectiveness and is applicable for a wide range of protein therapeutics and genetic diseases.

Highly accurate protein structure prediction with AlphaFold

Proteins are essential to life, and understanding their structure can facilitate a mechanistic understanding of their function. Through an enormous experimental effort1-4, the structures of around 100,000 unique proteins have been determined5, but this represents a small fraction of the billions of known protein sequences6,7. Structural coverage is bottlenecked by the months to years of painstaking effort required to determine a single protein structure. Accurate computational approaches are needed to address this gap and to enable large-scale structural bioinformatics. Predicting the three-dimensional structure that a protein will adopt based solely on its amino acid sequence-the structure prediction component of the 'protein folding problem'8-has been an important open research problem for more than 50 years9. Despite recent progress10-14, existing methods fall far short of atomic accuracy, especially when no homologous structure is available. Here we provide the first computational method that can regularly predict protein structures with atomic accuracy even in cases in which no similar structure is known. We validated an entirely redesigned version of our neural network-based model, AlphaFold, in the challenging 14th Critical Assessment of protein Structure Prediction (CASP14)15, demonstrating accuracy competitive with experimental structures in a majority of cases and greatly outperforming other methods. Underpinning the latest version of AlphaFold is a novel machine learning approach that incorporates physical and biological knowledge about protein structure, leveraging multi-sequence alignments, into the design of the deep learning algorithm.

Folding non-homologous proteins by coupling deep-learning contact maps with I-TASSER assembly simulations

Structure prediction for proteins lacking homologous templates in the Protein Data Bank (PDB) remains a significant unsolved problem. We developed a protocol, C-I-TASSER, to integrate interresidue contact maps from deep neural-network learning with the cutting-edge I-TASSER fragment assembly simulations. Large-scale benchmark tests showed that C-I-TASSER can fold more than twice the number of non-homologous proteins than the I-TASSER, which does not use contacts. When applied to a folding experiment on 8,266 unsolved Pfam families, C-I-TASSER successfully folded 4,162 domain families, including 504 folds that are not found in the PDB. Furthermore, it created correct folds for 85% of proteins in the SARS-CoV-2 genome, despite the quick mutation rate of the virus and sparse sequence profiles. The results demonstrated the critical importance of coupling whole-genome and metagenome-based evolutionary information with optimal structure assembly simulations for solving the problem of non-homologous protein structure prediction.

Comparative analysis of two paradigm bacteriophytochromes reveals opposite functionalities in two-component signaling

Bacterial phytochrome photoreceptors usually belong to two-component signaling systems which transmit environmental stimuli to a response regulator through a histidine kinase domain. Phytochromes switch between red light-absorbing and far-red light-absorbing states. Despite exhibiting extensive structural responses during this transition, the model bacteriophytochrome from Deinococcus radiodurans (DrBphP) lacks detectable kinase activity. Here, we resolve this long-standing conundrum by comparatively analyzing the interactions and output activities of DrBphP and a bacteriophytochrome from Agrobacterium fabrum (Agp1). Whereas Agp1 acts as a conventional histidine kinase, we identify DrBphP as a light-sensitive phosphatase. While Agp1 binds its cognate response regulator only transiently, DrBphP does so strongly, which is rationalized at the structural level. Our data pinpoint two key residues affecting the balance between kinase and phosphatase activities, which immediately bears on photoreception and two-component signaling. The opposing output activities in two highly similar bacteriophytochromes suggest the use of light-controllable histidine kinases and phosphatases for optogenetics.

TopDomain: Exhaustive Protein Domain Boundary Metaprediction Combining Multisource Information and Deep Learning

Protein domains are independent, functional, and stable structural units of proteins. Accurate protein domain boundary prediction plays an important role in understanding protein structure and evolution, as well as for protein structure prediction. Current domain boundary prediction methods differ in terms of boundary definition, methodology, and training databases resulting in disparate performance for different proteins. We developed TopDomain, an exhaustive metapredictor, that uses deep neural networks to combine multisource information from sequence- and homology-based features of over 50 primary predictors. For this purpose, we developed a new domain boundary data set termed the TopDomain data set, in which the true annotations are informed by SCOPe annotations, structural domain parsers, human inspection, and deep learning. We benchmark TopDomain against 2484 targets with 3354 boundaries from the TopDomain test set and achieve F1 scores of 78.4% and 73.8% for multidomain boundary prediction within ±20 residues and ±10 residues of the true boundary, respectively. When examined on targets from CASP11-13 competitions, TopDomain achieves F1 scores of 47.5% and 42.8% for multidomain proteins. TopDomain significantly outperforms 15 widely used, state-of-the-art ab initio and homology-based domain boundary predictors. Finally, we implemented TopDomain_TMC, which accurately predicts whether domain parsing is necessary for the target protein.

Obtaining extremely large and accurate protein multiple sequence alignments from curated hierarchical alignments

For optimal performance, machine learning methods for protein sequence/structural analysis typically require as input a large multiple sequence alignment (MSA), which is often created using query-based iterative programs, such as PSI-BLAST or JackHMMER. However, because these programs align database sequences using a query sequence as a template, they may fail to detect or may tend to misalign sequences distantly related to the query. More generally, automated MSA programs often fail to align sequences correctly due to the unpredictable nature of protein evolution. Addressing this problem typically requires manual curation in the light of structural data. However, curated MSAs tend to contain too few sequences to serve as input for statistically based methods. We address these shortcomings by making publicly available a set of 252 curated hierarchical MSAs (hiMSAs), containing a total of 26 212 066 sequences, along with programs for generating from these extremely large MSAs. Each hiMSA consists of a set of hierarchically arranged MSAs representing individual subgroups within a superfamily along with template MSAs specifying how to align each subgroup MSA against MSAs higher up the hierarchy. Central to this approach is the MAPGAPS search program, which uses a hiMSA as a query to align (potentially vast numbers of) matching database sequences with accuracy comparable to that of the curated hiMSA. We illustrate this process for the exonuclease–endonuclease–phosphatase superfamily and for pleckstrin homology domains. A set of extremely large MSAs generated from the hiMSAs in this way is available as input for deep learning, big data analyses. MAPGAPS, auxiliary programs CDD2MGS, AddPhylum, PurgeMSA and ConvertMSA and links to National Center for Biotechnology Information data files are available at https://www.igs.umaryland.edu/labs/neuwald/software/mapgaps/.

Toward the solution of the protein structure prediction problem

Since Anfinsen demonstrated that the information encoded in a protein's amino acid sequence determines its structure in 1973, solving the protein structure prediction problem has been the Holy Grail of structural biology. The goal of protein structure prediction approaches is to utilize computational modeling to determine the spatial location of every atom in a protein molecule starting from only its amino acid sequence. Depending on whether homologous structures can be found in the Protein Data Bank (PDB), structure prediction methods have been historically categorized as template-based modeling (TBM) or template-free modeling (FM) approaches. Until recently, TBM has been the most reliable approach to predicting protein structures, and in the absence of reliable templates, the modeling accuracy sharply declines. Nevertheless, the results of the most recent community-wide assessment of protein structure prediction experiment (CASP14) have demonstrated that the protein structure prediction problem can be largely solved through the use of end-to-end deep machine learning techniques, where correct folds could be built for nearly all single-domain proteins without using the PDB templates. Critically, the model quality exhibited little correlation with the quality of available template structures, as well as the number of sequence homologs detected for a given target protein. Thus, the implementation of deep-learning techniques has essentially broken through the 50-year-old modeling border between TBM and FM approaches and has made the success of high-resolution structure prediction significantly less dependent on template availability in the PDB library.

COMTOP: Protein Residue-Residue Contact Prediction through Mixed Integer Linear Optimization

Protein contact prediction helps reconstruct the tertiary structure that greatly determines a protein’s function; therefore, contact prediction from the sequence is an important problem. Recently there has been exciting progress on this problem, but many of the existing methods are still low quality of prediction accuracy. In this paper, we present a new mixed integer linear programming (MILP)-based consensus method: a Consensus scheme based On a Mixed integer linear opTimization method for prOtein contact Prediction (COMTOP). The MILP-based consensus method combines the strengths of seven selected protein contact prediction methods, including CCMpred, EVfold, DeepCov, NNcon, PconsC4, plmDCA, and PSICOV, by optimizing the number of correctly predicted contacts and achieving a better prediction accuracy. The proposed hybrid protein residue–residue contact prediction scheme was tested in four independent test sets. For 239 highly non-redundant proteins, the method showed a prediction accuracy of 59.68%, 70.79%, 78.86%, 89.04%, 94.51%, and 97.35% for top-5L, top-3L, top-2L, top-L, top-L/2, and top-L/5 contacts, respectively. When tested on the CASP13 and CASP14 test sets, the proposed method obtained accuracies of 75.91% and 77.49% for top-L/5 predictions, respectively. COMTOP was further tested on 57 non-redundant α-helical transmembrane proteins and achieved prediction accuracies of 64.34% and 73.91% for top-L/2 and top-L/5 predictions, respectively. For all test datasets, the improvement of COMTOP in accuracy over the seven individual methods increased with the increasing number of predicted contacts. For example, COMTOP performed much better for large number of contact predictions (such as top-5L and top-3L) than for small number of contact predictions such as top-L/2 and top-L/5. The results and analysis demonstrate that COMTOP can significantly improve the performance of the individual methods; therefore, COMTOP is more robust against different types of test sets. COMTOP also showed better/comparable predictions when compared with the state-of-the-art predictors.

Selective Nutrient Transport in Bacteria: Multicomponent Transporter Systems Reign Supreme

Multicomponent transporters are used by bacteria to transport a wide range of nutrients. These systems use a substrate-binding protein to bind the nutrient with high affinity and then deliver it to a membrane-bound transporter for uptake. Nutrient uptake pathways are linked to the colonisation potential and pathogenicity of bacteria in humans and may be candidates for antimicrobial targeting. Here we review current research into bacterial multicomponent transport systems, with an emphasis on the interaction at the membrane, as well as new perspectives on the role of lipids and higher oligomers in these complex systems.

CoeViz 2: Protein Graphs Derived From Amino Acid Covariance

Proteins by and large carry out their molecular functions in a folded state when residues, distant in sequence, assemble together in 3D space to bind a ligand, catalyze a reaction, form a channel, or exert another concerted macromolecular interaction. It has been long recognized that covariance of amino acids between distant positions within a protein sequence allows for the inference of long range contacts to facilitate 3D structure modeling. In this work, we investigated whether covariance analysis may reveal residues involved in the same molecular function. Building upon our previous work, CoeViz, we have conducted a large scale covariance analysis among 7,595 non-redundant proteins with resolved 3D structures to assess 1) whether the residues with the same function coevolve, 2) which covariance metric captures such couplings better, and 3) how different molecular functions compare in this context. We found that the chi-squared metric is the most informative for the identification of coevolving functional sites, followed by the Pearson correlation-based, whereas mutual information is the least informative. Of the seven categories of the most common natural ligands, including coenzyme A, dinucleotide, DNA/RNA, heme, metal, nucleoside, and sugar, the trace metal binding residues display the most prominent coupling, followed by the sugar binding sites. We also developed a web-based tool, CoeViz 2, that enables the interactive visualization of covarying residues as cliques from a larger protein graph. CoeViz 2 is publicly available at https://research.cchmc.org/CoevLab/.

Co-evolutionary landscape at the interface and non-interface regions of protein-protein interaction complexes

Proteins involved in interactions throughout the course of evolution tend to co-evolve and compensatory changes may occur in interacting proteins to maintain or refine such interactions. However, certain residue pair alterations may prove to be detrimental for functional interactions. Hence, determining co-evolutionary pairings that could be structurally or functionally relevant for maintaining the conservation of an inter-protein interaction is important. Inter-protein co-evolution analysis in several complexes utilizing multiple existing methodologies suggested that co-evolutionary pairings can occur in spatially proximal and distant regions in inter-protein interactions. Subsequently, the Co-Var (Correlated Variation) method based on mutual information and Bhattacharyya coefficient was developed, validated, and found to perform relatively better than CAPS and EV-complex. Interestingly, while applying the Co-Var measure and EV-complex program on a set of protein-protein interaction complexes, co-evolutionary pairings were obtained in interface and non-interface regions in protein complexes. The Co-Var approach involves determining high degree co-evolutionary pairings that include multiple co-evolutionary connections between particular co-evolved residue positions in one protein with multiple residue positions in the binding partner. Detailed analyses of high degree co-evolutionary pairings in protein-protein complexes involved in cancer metastasis suggested that most of the residue positions forming such co-evolutionary connections mainly occurred within functional domains of constituent proteins and substitution mutations were also common among these positions. The physiological relevance of these predictions suggested that Co-Var can predict residues that could be crucial for preserving functional protein-protein interactions. Finally, Co-Var web server (http://www.hpppi.iicb.res.in/ishi/covar/index.html) that implements this methodology identifies co-evolutionary pairings in intra and inter-protein interactions.

DNCON2_Inter: predicting interchain contacts for homodimeric and homomultimeric protein complexes using multiple sequence alignments of monomers and deep learning

Deep learning methods that achieved great success in predicting intrachain residue-residue contacts have been applied to predict interchain contacts between proteins. However, these methods require multiple sequence alignments (MSAs) of a pair of interacting proteins (dimers) as input, which are often difficult to obtain because there are not many known protein complexes available to generate MSAs of sufficient depth for a pair of proteins. In recognizing that multiple sequence alignments of a monomer that forms homomultimers contain the co-evolutionary signals of both intrachain and interchain residue pairs in contact, we applied DNCON2 (a deep learning-based protein intrachain residue-residue contact predictor) to predict both intrachain and interchain contacts for homomultimers using multiple sequence alignment (MSA) and other co-evolutionary features of a single monomer followed by discrimination of interchain and intrachain contacts according to the tertiary structure of the monomer. We name this tool DNCON2_Inter. Allowing true-positive predictions within two residue shifts, the best average precision was obtained for the Top-L/10 predictions of 22.9% for homodimers and 17.0% for higher-order homomultimers. In some instances, especially where interchain contact densities are high, DNCON2_Inter predicted interchain contacts with 100% precision. We also developed Con_Complex, a complex structure reconstruction tool that uses predicted contacts to produce the structure of the complex. Using Con_Complex, we show that the predicted contacts can be used to accurately construct the structure of some complexes. Our experiment demonstrates that monomeric multiple sequence alignments can be used with deep learning to predict interchain contacts of homomeric proteins.

DELPHI: accurate deep ensemble model for protein interaction sites prediction

MOTIVATION

Proteins usually perform their functions by interacting with other proteins, which is why accurately predicting protein-protein interaction (PPI) binding sites is a fundamental problem. Experimental methods are slow and expensive. Therefore, great efforts are being made towards increasing the performance of computational methods.

RESULTS

We propose DEep Learning Prediction of Highly probable protein Interaction sites (DELPHI), a new sequence-based deep learning suite for PPI-binding sites prediction. DELPHI has an ensemble structure which combines a CNN and a RNN component with fine tuning technique. Three novel features, HSP, position information and ProtVec are used in addition to nine existing ones. We comprehensively compare DELPHI to nine state-of-the-art programmes on five datasets, and DELPHI outperforms the competing methods in all metrics even though its training dataset shares the least similarities with the testing datasets. In the most important metrics, AUPRC and MCC, it surpasses the second best programmes by as much as 18.5% and 27.7%, respectively. We also demonstrated that the improvement is essentially due to using the ensemble model and, especially, the three new features. Using DELPHI it is shown that there is a strong correlation with protein-binding residues (PBRs) and sites with strong evolutionary conservation. In addition, DELPHI's predicted PBR sites closely match known data from Pfam. DELPHI is available as open-sourced standalone software and web server.

AVAILABILITY AND IMPLEMENTATION

The DELPHI web server can be found at delphi.csd.uwo.ca/, with all datasets and results in this study. The trained models, the DELPHI standalone source code, and the feature computation pipeline are freely available at github.com/lucian-ilie/DELPHI.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Structural discrimination analysis for constraint selection in protein modeling

MOTIVATION

Protein structure modeling can be improved by the use of distance constraints between amino acid residues, provided such data reflects-at least partially-the native tertiary structure of the target system. In fact, only a small subset of the native contact map is necessary to successfully drive the model conformational search, so one important goal is to obtain the set of constraints with the highest true-positive rate, lowest redundancy, and greatest amount of information. In this work, we introduce a constraint evaluation and selection method based on the point-biserial correlation coefficient, which utilizes structural information from an ensemble of models to indirectly measure the power of each constraint in biasing the conformational search towards consensus structures.

RESULTS

Residue contact maps obtained by direct coupling analysis are systematically improved by means of discriminant analysis, reaching in some cases accuracies often seen only in modern deep-learning based approaches. When combined with an iterative modeling workflow, the proposed constraint classification optimizes the selection of the constraint set and maximizes the probability of obtaining successful models. The use of discriminant analysis for the valorization of the information of constraint data sets is a general concept with possible applications to other constraint types and modeling problems.

AVAILABILITY AND IMPLEMENTATION

scripts and procedures to implement the methodology presented herein are available at https://github.com/m3g/2021_Bottino_Biserial.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Recent Applications of Deep Learning Methods on Evolution- and Contact-Based Protein Structure Prediction

The new advances in deep learning methods have influenced many aspects of scientific research, including the study of the protein system. The prediction of proteins' 3D structural components is now heavily dependent on machine learning techniques that interpret how protein sequences and their homology govern the inter-residue contacts and structural organization. Especially, methods employing deep neural networks have had a significant impact on recent CASP13 and CASP14 competition. Here, we explore the recent applications of deep learning methods in the protein structure prediction area. We also look at the potential opportunities for deep learning methods to identify unknown protein structures and functions to be discovered and help guide drug-target interactions. Although significant problems still need to be addressed, we expect these techniques in the near future to play crucial roles in protein structural bioinformatics as well as in drug discovery.

Deep learning techniques have significantly impacted protein structure prediction and protein design

Protein structure prediction and design can be regarded as two inverse processes governed by the same folding principle. Although progress remained stagnant over the past two decades, the recent application of deep neural networks to spatial constraint prediction and end-to-end model training has significantly improved the accuracy of protein structure prediction, largely solving the problem at the fold level for single-domain proteins. The field of protein design has also witnessed dramatic improvement, where noticeable examples have shown that information stored in neural-network models can be used to advance functional protein design. Thus, incorporation of deep learning techniques into different steps of protein folding and design approaches represents an exciting future direction and should continue to have a transformative impact on both fields.

Current directions in combining simulation-based macromolecular modeling approaches with deep learning

Introduction: Structure-guided drug discovery relies on accurate computational methods for modeling macromolecules. Simulations provide means of predicting macromolecular folds, of discovering function from structure, and of designing macromolecules to serve as drugs. Success rates are limited for any of these tasks, however. Recently, deep neural network-based methods have greatly enhanced the accuracy of predictions of protein structure from sequence, generating excitement about the potential impact of deep learning.Areas covered: This review introduces biologists to deep neural network architecture, surveys recent successes of deep learning in structure prediction, and discusses emerging deep learning-based approaches for structure-function analysis and design. Particular focus is given to the interplay between simulation-based and neural network-based approaches.Expert opinion: As deep learning grows integral to macromolecular modeling, simulation- and neural network-based approaches must grow more tightly interconnected. Modular software architecture must emerge allowing both types of tools to be combined with maximal versatility. Open sharing of code under permissive licenses will be essential. Although experiments will remain the gold standard for reliable information to guide drug discovery, we may soon see successful drug development projects based on high-accuracy predictions from algorithms that combine simulation with deep learning - the ultimate validation of this combination's power.

On the effect of phylogenetic correlations in coevolution-based contact prediction in proteins

Coevolution-based contact prediction, either directly by coevolutionary couplings resulting from global statistical sequence models or using structural supervision and deep learning, has found widespread application in protein-structure prediction from sequence. However, one of the basic assumptions in global statistical modeling is that sequences form an at least approximately independent sample of an unknown probability distribution, which is to be learned from data. In the case of protein families, this assumption is obviously violated by phylogenetic relations between protein sequences. It has turned out to be notoriously difficult to take phylogenetic correlations into account in coevolutionary model learning. Here, we propose a complementary approach: we develop strategies to randomize or resample sequence data, such that conservation patterns and phylogenetic relations are preserved, while intrinsic (i.e. structure- or function-based) coevolutionary couplings are removed. A comparison between the results of Direct Coupling Analysis applied to real and to resampled data shows that the largest coevolutionary couplings, i.e. those used for contact prediction, are only weakly influenced by phylogeny. However, the phylogeny-induced spurious couplings in the resampled data are compatible in size with the first false-positive contact predictions from real data. Dissecting functional from phylogeny-induced couplings might therefore extend accurate contact predictions to the range of intermediate-size couplings.

Many homologous protein families contain thousands of highly diverged amino-acid sequences, which fold into close-to-identical three-dimensional structures and fulfill almost identical biological tasks. Global coevolutionary models, like those inferred by the Direct Coupling Analysis (DCA), assume that families can be considered as samples of some unknown statistical model, and that the parameters of these models represent evolutionary constraints acting on protein sequences. To learn these models from data, DCA and related approaches have to also assume that the distinct sequences in a protein family are close to independent, while in reality they are characterized by involved hierarchical phylogenetic relationships. Here we propose Null models for sequence alignments, which maintain patterns of amino-acid conservation and phylogeny contained in the data, but destroy any coevolutionary couplings, frequently used in protein structure prediction. We find that phylogeny actually induces spurious non-zero couplings. These are, however, significantly smaller that the largest couplings derived from natural sequences, and therefore have only little influence on the first predicted contacts. However, in the range of intermediate couplings, they may lead to statistically significant effects. Dissecting phylogenetic from functional couplings might therefore extend the range of accurately predicted structural contacts down to smaller coupling strengths than those currently used.

Evaluation of residue-residue contact prediction methods: From retrospective to prospective

Sequence-based residue contact prediction plays a crucial role in protein structure reconstruction. In recent years, the combination of evolutionary coupling analysis (ECA) and deep learning (DL) techniques has made tremendous progress for residue contact prediction, thus a comprehensive assessment of current methods based on a large-scale benchmark data set is very needed. In this study, we evaluate 18 contact predictors on 610 non-redundant proteins and 32 CASP13 targets according to a wide range of perspectives. The results show that different methods have different application scenarios: (1) DL methods based on multi-categories of inputs and large training sets are the best choices for low-contact-density proteins such as the intrinsically disordered ones and proteins with shallow multi-sequence alignments (MSAs). (2) With at least 5L (L is sequence length) effective sequences in the MSA, all the methods show the best performance, and methods that rely only on MSA as input can reach comparable achievements as methods that adopt multi-source inputs. (3) For top L/5 and L/2 predictions, DL methods can predict more hydrophobic interactions while ECA methods predict more salt bridges and disulfide bonds. (4) ECA methods can detect more secondary structure interactions, while DL methods can accurately excavate more contact patterns and prune isolated false positives. In general, multi-input DL methods with large training sets dominate current approaches with the best overall performance. Despite the great success of current DL methods must be stated the fact that there is still much room left for further improvement: (1) With shallow MSAs, the performance will be greatly affected. (2) Current methods show lower precisions for inter-domain compared with intra-domain contact predictions, as well as very high imbalances in precisions between intra-domains. (3) Strong prediction similarities between DL methods indicating more feature types and diversified models need to be developed. (4) The runtime of most methods can be further optimized.

The amino acid sequence of a protein ultimately determines its tertiary structure, and the tertiary structure determines its function(s) and plays a key role in understanding biological processes and disease pathogenesis. Protein tertiary structure can be determined using experimental techniques such as cryo-electron microscopy, nuclear magnetic resonance and X-ray crystallography, which are very expensive and time-consuming. As an alternative, researchers are trying to use in silico methods to predict the 3D structures. Residue contact-assisted protein folding paves an avenue for sequence-based protein structure prediction and therefore has become one of the most challenging and promising problems in structural bioinformatics. Over the past years, contact prediction has undergone continuous evolution in techniques. Through a retrospective analysis of traditional machine learning /evolutionary coupling analysis methods/ consensus machine learning methods and a multi-perspective study on recently developed deep learning methods, we explore the most advanced contact predictors, pursue application scenarios for different methods, and seek prospective directions for further improvement. We anticipate that our study will serve as a practical and useful guide for the development of future approaches to contact prediction.

RNAcmap: A Fully Automatic Pipeline for Predicting Contact Maps of RNAs by Evolutionary Coupling Analysis

MOTIVATION

The accuracy of RNA secondary and tertiary structure prediction can be significantly improved by using structural restraints derived from evolutionary coupling or direct coupling analysis. Currently, these coupling analyses relied on manually curated multiple sequence alignments collected in the Rfam database, which contains 3016 families. By comparison, millions of non-coding RNA sequences are known. Here, we established RNAcmap, a fully automatic pipeline that enables evolutionary coupling analysis for any RNA sequences. The homology search was based on the covariance model built by INFERNAL according to two secondary structure predictors: a folding-based algorithm RNAfold and the latest deep-learning method SPOT-RNA.

RESULTS

We showed that the performance of RNAcmap is less dependent on the specific evolutionary coupling tool but is more dependent on the accuracy of secondary structure predictor with the best performance given by RNAcmap (SPOT-RNA). The performance of RNAcmap (SPOT-RNA) is comparable to that based on Rfam-supplied alignment and consistent for those sequences that are not in Rfam collections. Further improvement can be made with a simple meta predictor RNAcmap (SPOT-RNA/RNAfold) depending on which secondary structure predictor can find more homologous sequences. Reliable base-pairing information generated from RNAcmap, for RNAs with high effective homologous sequences, in particular, will be useful for aiding RNA structure prediction.

AVAILABILITY

RNAcmap is available as a web server at https://sparks-lab.org/server/rnacmap/ and as a standalone application along with the datasets at https://github.com/sparks-lab-org/RNAcmap_standalone. A platform independent and fully configured docker image of RNAcmap is also provided at https://hub.docker.com/r/jaswindersingh2/rnacmap.

AlphaFold and the amyloid landscape

Protein aggregation is a widespread phenomenon with important implications in many scientific areas. Although amyloid formation is typically considered as detrimental, functional amyloids that perform physiological roles have been identified in all kingdoms of life. Despite their functional and pathological relevance, the structural details of the majority of molecular species involved in the amyloidogenic process remains elusive. Here, we explore the application of AlphaFold, a highly accurate protein structure predictor, in the field of protein aggregation. While we envision a straightforward application of AlphaFold in assisting the design of globular proteins with improved solubility for biomedical and industrial purposes, the use of this algorithm for predicting the structure of aggregated species seems far from trivial. First, in amyloid diseases, the presence of multiple amyloid polymorphs and the heterogeneity of aggregation intermediates challenges the "one sequence, one structure" paradigm, inherent to sequence-based predictions. Second, aberrant aggregation is not the subject of positive selective pressure, precluding the use of evolutionary-based approaches, which are the core of the AlphaFold pipeline. Instead, amyloid polymorphism seems to be constrained by the need for a defined structure-activity relationship in functional amyloids. They may thus provide a starting point for the application of AlphaFold in the amyloid landscape.

Recent advances in user-friendly computational tools to engineer protein function

Progress in technology and algorithms throughout the past decade has transformed the field of protein design and engineering. Computational approaches have become well-engrained in the processes of tailoring proteins for various biotechnological applications. Many tools and methods are developed and upgraded each year to satisfy the increasing demands and challenges of protein engineering. To help protein engineers and bioinformaticians navigate this emerging wave of dedicated software, we have critically evaluated recent additions to the toolbox regarding their application for semi-rational and rational protein engineering. These newly developed tools identify and prioritize hotspots and analyze the effects of mutations for a variety of properties, comprising ligand binding, protein-protein and protein-nucleic acid interactions, and electrostatic potential. We also discuss notable progress to target elusive protein dynamics and associated properties like ligand-transport processes and allosteric communication. Finally, we discuss several challenges these tools face and provide our perspectives on the further development of readily applicable methods to guide protein engineering efforts.

Identification of biophysical interaction patterns in direct coupling analysis

Direct-coupling analysis is a statistical learning method for protein contact prediction based on sequence information alone. The maximum entropy principle leads to an effective inverse Potts model. Predictions on contacts are based on fitted local fields and couplings from an empirical multiple sequence alignment. Typically, the l_{2} norm of the resulting two-body couplings is used for contact prediction. However, this procedure discards important information. In this paper we show that the usage of the full fields and coupling information improves prediction accuracy.

Machine learning in protein structure prediction

Prediction of protein structure from sequence has been intensely studied for many decades, owing to the problem's importance and its uniquely well-defined physical and computational bases. While progress has historically ebbed and flowed, the past two years saw dramatic advances driven by the increasing "neuralization" of structure prediction pipelines, whereby computations previously based on energy models and sampling procedures are replaced by neural networks. The extraction of physical contacts from the evolutionary record; the distillation of sequence-structure patterns from known structures; the incorporation of templates from homologs in the Protein Databank; and the refinement of coarsely predicted structures into finely resolved ones have all been reformulated using neural networks. Cumulatively, this transformation has resulted in algorithms that can now predict single protein domains with a median accuracy of 2.1 Å, setting the stage for a foundational reconfiguration of the role of biomolecular modeling within the life sciences.

eCOMPASS: evaluative comparison of multiple protein alignments by statistical score

Motivation

Detecting subtle biologically relevant patterns in protein sequences often requires the construction of a large and accurate multiple sequence alignment (MSA). Methods for constructing MSAs are usually evaluated using benchmark alignments, which, however, typically contain very few sequences and are therefore inappropriate when dealing with large numbers of proteins.

Results

eCOMPASS addresses this problem using a statistical measure of relative alignment quality based on direct coupling analysis (DCA): to maintain protein structural integrity over evolutionary time, substitutions at one residue position typically result in compensating substitutions at other positions. eCOMPASS computes the statistical significance of the congruence between high scoring directly coupled pairs and 3D contacts in corresponding structures, which depends upon properly aligned homologous residues. We illustrate eCOMPASS using both simulated and real MSAs.

Availability and implementation

The eCOMPASS executable, C++ open source code and input data sets are available at https://www.igs.umaryland.edu/labs/neuwald/software/compass

Supplementary information

Supplementary data are available at Bioinformatics online.

CopulaNet: Learning residue co-evolution directly from multiple sequence alignment for protein structure prediction

Residue co-evolution has become the primary principle for estimating inter-residue distances of a protein, which are crucially important for predicting protein structure. Most existing approaches adopt an indirect strategy, i.e., inferring residue co-evolution based on some hand-crafted features, say, a covariance matrix, calculated from multiple sequence alignment (MSA) of target protein. This indirect strategy, however, cannot fully exploit the information carried by MSA. Here, we report an end-to-end deep neural network, CopulaNet, to estimate residue co-evolution directly from MSA. The key elements of CopulaNet include: (i) an encoder to model context-specific mutation for each residue; (ii) an aggregator to model residue co-evolution, and thereafter estimate inter-residue distances. Using CASP13 (the 13th Critical Assessment of Protein Structure Prediction) target proteins as representatives, we demonstrate that CopulaNet can predict protein structure with improved accuracy and efficiency. This study represents a step toward improved end-to-end prediction of inter-residue distances and protein tertiary structures.

Extending the New Generation of Structure Predictors to Account for Dynamics and Allostery

Recent progress in structure-prediction methods that rely on deep learning suggests that the atomic structure of almost any protein may soon be predictable directly from its amino acid sequence. This much-awaited revolution was driven by substantial improvements in the reliability of methods for inferring the spatial distances between amino acid pairs from an analysis of homologous sequences. Improved reliability has been accompanied, however, by a reduced ability to detect amino acid relationships that are not due to direct spatial contacts, such as those that arise from protein dynamics or allostery. Given the central importance of dynamics and allostery to protein activity, we argue that an important future advance would extend modeling beyond predicting a single static structure. Here, we briefly review some of the developments that have led to the remarkable recent achievement in structure prediction and speculate what methods and sources of information may be leveraged in the future to develop a modeling framework that addresses protein dynamics and allostery.

Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences

In the field of artificial intelligence, a combination of scale in data and model capacity enabled by unsupervised learning has led to major advances in representation learning and statistical generation. In the life sciences, the anticipated growth of sequencing promises unprecedented data on natural sequence diversity. Protein language modeling at the scale of evolution is a logical step toward predictive and generative artificial intelligence for biology. To this end, we use unsupervised learning to train a deep contextual language model on 86 billion amino acids across 250 million protein sequences spanning evolutionary diversity. The resulting model contains information about biological properties in its representations. The representations are learned from sequence data alone. The learned representation space has a multiscale organization reflecting structure from the level of biochemical properties of amino acids to remote homology of proteins. Information about secondary and tertiary structure is encoded in the representations and can be identified by linear projections. Representation learning produces features that generalize across a range of applications, enabling state-of-the-art supervised prediction of mutational effect and secondary structure and improving state-of-the-art features for long-range contact prediction.

Analyzing effect of quadruple multiple sequence alignments on deep learning based protein inter-residue distance prediction

Protein 3D structure prediction has advanced significantly in recent years due to improving contact prediction accuracy. This improvement has been largely due to deep learning approaches that predict inter-residue contacts and, more recently, distances using multiple sequence alignments (MSAs). In this work we present AttentiveDist, a novel approach that uses different MSAs generated with different E-values in a single model to increase the co-evolutionary information provided to the model. To determine the importance of each MSA's feature at the inter-residue level, we added an attention layer to the deep neural network. We show that combining four MSAs of different E-value cutoffs improved the model prediction performance as compared to single E-value MSA features. A further improvement was observed when an attention layer was used and even more when additional prediction tasks of bond angle predictions were added. The improvement of distance predictions were successfully transferred to achieve better protein tertiary structure modeling.

Deducing high-accuracy protein contact-maps from a triplet of coevolutionary matrices through deep residual convolutional networks

The topology of protein folds can be specified by the inter-residue contact-maps and accurate contact-map prediction can help ab initio structure folding. We developed TripletRes to deduce protein contact-maps from discretized distance profiles by end-to-end training of deep residual neural-networks. Compared to previous approaches, the major advantage of TripletRes is in its ability to learn and directly fuse a triplet of coevolutionary matrices extracted from the whole-genome and metagenome databases and therefore minimize the information loss during the course of contact model training. TripletRes was tested on a large set of 245 non-homologous proteins from CASP 11&12 and CAMEO experiments and outperformed other top methods from CASP12 by at least 58.4% for the CASP 11&12 targets and 44.4% for the CAMEO targets in the top-L long-range contact precision. On the 31 FM targets from the latest CASP13 challenge, TripletRes achieved the highest precision (71.6%) for the top-L/5 long-range contact predictions. It was also shown that a simple re-training of the TripletRes model with more proteins can lead to further improvement with precisions comparable to state-of-the-art methods developed after CASP13. These results demonstrate a novel efficient approach to extend the power of deep convolutional networks for high-accuracy medium- and long-range protein contact-map predictions starting from primary sequences, which are critical for constructing 3D structure of proteins that lack homologous templates in the PDB library.

Ab initio protein folding has been a major unsolved problem in computational biology for more than half a century. Recent community-wide Critical Assessment of Structure Prediction (CASP) experiments have witnessed exciting progress on ab initio structure prediction, which was mainly powered by the boosting of contact-map prediction as the latter can be used as constraints to guide ab initio folding simulations. In this work, we proposed a new open-source deep-learning architecture, TripletRes, built on the residual convolutional neural networks for high-accuracy contact prediction. The large-scale benchmark and blind test results demonstrate competitive performance of the proposed methods to other top approaches in predicting medium- and long-range contact-maps that are critical for guiding protein folding simulations. Detailed data analyses showed that the major advantage of TripletRes lies in the unique protocol to fuse multiple evolutionary feature matrices which are directly extracted from whole-genome and metagenome databases and therefore minimize the information loss during the contact model training.

Evolutionary conservation of RNA sequence and structure

An RNA structure prediction from a single‐sequence RNA folding program is not evidence for an RNA whose structure is important for function. Random sequences have plausible and complex predicted structures not easily distinguishable from those of structural RNAs. How to tell when an RNA has a conserved structure is a question that requires looking at the evolutionary signature left by the conserved RNA. This question is important not just for long noncoding RNAs which usually lack an identified function, but also for RNA binding protein motifs which can be single stranded RNAs or structures. Here we review recent advances using sequence and structural analysis to determine when RNA structure is conserved or not. Although covariation measures assess structural RNA conservation, one must distinguish covariation due to RNA structure from covariation due to independent phylogenetic substitutions. We review a statistical test to measure false positives expected under the null hypothesis of phylogenetic covariation alone (specificity). We also review a complementary test that measures power, that is, expected covariation derived from sequence variation alone (sensitivity). Power in the absence of covariation signals the absence of a conserved RNA structure. We analyze artifacts that falsely identify conserved RNA structure such as the misuse of programs that do not assess significance, the use of inappropriate statistics confounded by signals other than covariation, or misalignments that induce spurious covariation. Among artifacts that obscure the signal of a conserved RNA structure, we discuss the inclusion of pseudogenes in alignments which increase power but destroy covariation.

This article is categorized under:

RNA Structure and Dynamics > RNA Structure, Dynamics and Chemistry

RNA Evolution and Genomics > Computational Analyses of RNA

RNA Evolution and Genomics > RNA and Ribonucleoprotein Evolution

Combination of deep neural network with attention mechanism enhances the explainability of protein contact prediction

Deep learning has emerged as a revolutionary technology for protein residue‐residue contact prediction since the 2012 CASP10 competition. Considerable advancements in the predictive power of the deep learning‐based contact predictions have been achieved since then. However, little effort has been put into interpreting the black‐box deep learning methods. Algorithms that can interpret the relationship between predicted contact maps and the internal mechanism of the deep learning architectures are needed to explore the essential components of contact inference and improve their explainability. In this study, we present an attention‐based convolutional neural network for protein contact prediction, which consists of two attention mechanism‐based modules: sequence attention and regional attention. Our benchmark results on the CASP13 free‐modeling targets demonstrate that the two attention modules added on top of existing typical deep learning models exhibit a complementary effect that contributes to prediction improvements. More importantly, the inclusion of the attention mechanism provides interpretable patterns that contain useful insights into the key fold‐determining residues in proteins. We expect the attention‐based model can provide a reliable and practically interpretable technique that helps break the current bottlenecks in explaining deep neural networks for contact prediction. The source code of our method is available at https://github.com/jianlin-cheng/InterpretContactMap.

NGlyAlign: an automated library building tool to align highly divergent HIV envelope sequences

BACKGROUND

The high variability in envelope regions of some viruses such as HIV allow the virus to establish infection and to escape subsequent immune surveillance. This variability, as well as increasing incorporation of N-linked glycosylation sites, is fundamental to this evasion. It also creates difficulties for multiple sequence alignment methods (MSA) that provide the first step in their analysis. Existing MSA tools often fail to properly align highly variable HIV envelope sequences requiring extensive manual editing that is impractical with even a moderate number of these variable sequences.

RESULTS

We developed an automated library building tool NGlyAlign, that organizes similar N-linked glycosylation sites as block constraints and statistically conserved global sites as single site constraints to automatically enforce partial columns in consistency-based MSA methods such as Dialign. This combined method accurately aligns variable HIV-1 envelope sequences. We tested the method on two datasets: a set of 156 founder and chronic gp160 HIV-1 subtype B sequences as well as a set of reference sequences of gp120 in the highly variable region 1. On measures such as entropy scores, sum of pair scores, column score, and similarity heat maps, NGlyAlign+Dialign proved superior against methods such as T-Coffee, ClustalOmega, ClustalW, Praline, HIValign and Muscle. The method is scalable to large sequence sets producing accurate alignments without requiring manual editing. As well as this application to HIV, our method can be used for other highly variable glycoproteins such as hepatitis C virus envelope.

CONCLUSIONS

NGlyAlign is an automated tool for mapping and building glycosylation motif libraries to accurately align highly variable regions in HIV sequences. It can provide the basis for many studies reliant on single robust alignments. NGlyAlign has been developed as an open-source tool and is freely available at https://github.com/UNSW-Mathematical-Biology/NGlyAlign_v1.0 .

Why can deep convolutional neural networks improve protein fold recognition? A visual explanation by interpretation

As an essential task in protein structure and function prediction, protein fold recognition has attracted increasing attention. The majority of the existing machine learning-based protein fold recognition approaches strongly rely on handcrafted features, which depict the characteristics of different protein folds; however, effective feature extraction methods still represent the bottleneck for further performance improvement of protein fold recognition. As a powerful feature extractor, deep convolutional neural network (DCNN) can automatically extract discriminative features for fold recognition without human intervention, which has demonstrated an impressive performance on protein fold recognition. Despite the encouraging progress, DCNN often acts as a black box, and as such, it is challenging for users to understand what really happens in DCNN and why it works well for protein fold recognition. In this study, we explore the intrinsic mechanism of DCNN and explain why it works for protein fold recognition using a visual explanation technique. More specifically, we first trained a VGGNet-based DCNN model, termed VGGNet-FE, which can extract fold-specific features from the predicted protein residue-residue contact map for protein fold recognition. Subsequently, based on the trained VGGNet-FE, we implemented a new contact-assisted predictor, termed VGGfold, for protein fold recognition; we then visualized what features were extracted by each of the convolutional layers in VGGNet-FE using a deconvolution technique. Furthermore, we visualized the high-level semantic information, termed fold-discriminative region, of a predicted contact map from the localization map obtained from the last convolutional layer of VGGNet-FE. It is visually confirmed that VGGNet-FE could effectively extract distinct fold-discriminative regions for different types of protein folds, thereby accounting for the improved performance of VGGfold for protein fold recognition. In summary, this study is of great significance for both understanding the working principle of DCNNs in protein fold recognition and exploring the relationship between the predicted protein contact map and protein tertiary structure. This proposed visualization method is flexible and applicable to address other DCNN-based bioinformatics and computational biology questions. The online web server of VGGfold is freely available at http://csbio.njust.edu.cn/bioinf/vggfold/.

Generating functional protein variants with variational autoencoders

The vast expansion of protein sequence databases provides an opportunity for new protein design approaches which seek to learn the sequence-function relationship directly from natural sequence variation. Deep generative models trained on protein sequence data have been shown to learn biologically meaningful representations helpful for a variety of downstream tasks, but their potential for direct use in the design of novel proteins remains largely unexplored. Here we show that variational autoencoders trained on a dataset of almost 70000 luciferase-like oxidoreductases can be used to generate novel, functional variants of the luxA bacterial luciferase. We propose separate VAE models to work with aligned sequence input (MSA VAE) and raw sequence input (AR-VAE), and offer evidence that while both are able to reproduce patterns of amino acid usage characteristic of the family, the MSA VAE is better able to capture long-distance dependencies reflecting the influence of 3D structure. To confirm the practical utility of the models, we used them to generate variants of luxA whose luminescence activity was validated experimentally. We further showed that conditional variants of both models could be used to increase the solubility of luxA without disrupting function. Altogether 6/12 of the variants generated using the unconditional AR-VAE and 9/11 generated using the unconditional MSA VAE retained measurable luminescence, together with all 23 of the less distant variants generated by conditional versions of the models; the most distant functional variant contained 35 differences relative to the nearest training set sequence. These results demonstrate the feasibility of using deep generative models to explore the space of possible protein sequences and generate useful variants, providing a method complementary to rational design and directed evolution approaches.

DeepDist: real-value inter-residue distance prediction with deep residual convolutional network

BACKGROUND

Driven by deep learning, inter-residue contact/distance prediction has been significantly improved and substantially enhanced ab initio protein structure prediction. Currently, most of the distance prediction methods classify inter-residue distances into multiple distance intervals instead of directly predicting real-value distances. The output of the former has to be converted into real-value distances to be used in tertiary structure prediction.

RESULTS

To explore the potentials of predicting real-value inter-residue distances, we develop a multi-task deep learning distance predictor (DeepDist) based on new residual convolutional network architectures to simultaneously predict real-value inter-residue distances and classify them into multiple distance intervals. Tested on 43 CASP13 hard domains, DeepDist achieves comparable performance in real-value distance prediction and multi-class distance prediction. The average mean square error (MSE) of DeepDist's real-value distance prediction is 0.896 Ų when filtering out the predicted distance ≥ 16 Å, which is lower than 1.003 Ų of DeepDist's multi-class distance prediction. When distance predictions are converted into contact predictions at 8 Å threshold (the standard threshold in the field), the precision of top L/5 and L/2 contact predictions of DeepDist's multi-class distance prediction is 79.3% and 66.1%, respectively, higher than 78.6% and 64.5% of its real-value distance prediction and the best results in the CASP13 experiment.

CONCLUSIONS

DeepDist can predict inter-residue distances well and improve binary contact prediction over the existing state-of-the-art methods. Moreover, the predicted real-value distances can be directly used to reconstruct protein tertiary structures better than multi-class distance predictions due to the lower MSE. Finally, we demonstrate that predicting the real-value distance map and multi-class distance map at the same time performs better than predicting real-value distances alone.

Episodic evolution of coadapted sets of amino acid sites in mitochondrial proteins

The rate of evolution differs between protein sites and changes with time. However, the link between these two phenomena remains poorly understood. Here, we design a phylogenetic approach for distinguishing pairs of amino acid sites that evolve concordantly, i.e., such that substitutions at one site trigger subsequent substitutions at the other; and also pairs of sites that evolve discordantly, so that substitutions at one site impede subsequent substitutions at the other. We distinguish groups of amino acid sites that undergo coordinated evolution and evolve discordantly from other such groups. In mitochondrion-encoded proteins of metazoans and fungi, we show that concordantly evolving sites are clustered in protein structures. By analysing the phylogenetic patterns of substitutions at concordantly and discordantly evolving site pairs, we find that concordant evolution has two distinct causes: epistatic interactions between amino acid substitutions and episodes of selection independently affecting substitutions at different sites. The rate of substitutions at concordantly evolving groups of protein sites changes in the course of evolution, indicating episodes of selection limited to some of the lineages. The phylogenetic positions of these changes are consistent between proteins, suggesting common selective forces underlying them.

The mode and rate of evolution of a protein site depends on the effect of its mutations on protein fitness. The fitness effect of a mutation itself can change in the course of evolution for at least two reasons. First, it can be modulated by substitutions occurring at other sites, a phenomenon called epistasis. Second, changes in selection can be non-epistatic, affecting sites independently of one another. Here, we analyse substitutions accumulated by the evolving lineages of the five proteins encoded by the mitochondrial genomes of thousands of species of metazoans and fungi. We show that substitutions at different amino acid sites occur in a coordinated fashion, and this coordination is caused both by epistasis and by episodes of selection affecting groups of sites. We partition each protein into several groups of concordantly evolving sites such that evolution of sites from different groups is discordant, and show that the proteins encoded by the mitochondrial genome consist of coevolving structural blocks. Some of these blocks have a clear functional specialization, e.g. are associated with interfaces between proteins composing respiratory complexes. Together, our results reveal a previously unrecognized complexity in the causes of variation in evolutionary rates between protein sites.

New amino acid substitution matrix brings sequence alignments into agreement with structure matches

Protein sequence matching presently fails to identify many structures that are highly similar, even when they are known to have the same function. The high packing densities in globular proteins lead to interdependent substitutions, which have not previously been considered for amino acid similarities. At present, sequence matching compares sequences based only upon the similarities of single amino acids, ignoring the fact that in densely packed protein, there are additional conservative substitutions representing exchanges between two interacting amino acids, such as a small-large pair changing to a large-small pair substitutions that are not individually so conservative. Here we show that including information for such pairs of substitutions yields improved sequence matches, and that these yield significant gains in the agreements between sequence alignments and structure matches of the same protein pair. The result shows sequence segments matched where structure segments are aligned. There are gains for all 2002 collected cases where the sequence alignments that were not previously congruent with the structure matches. Our results also demonstrate a significant gain in detecting homology for “twilight zone” protein sequences. The amino acid substitution metrics derived have many other potential applications, for annotations, protein design, mutagenesis design, and empirical potential derivation.

Rosetta design with co-evolutionary information retains protein function

Computational protein design has the ambitious goal of crafting novel proteins that address challenges in biology and medicine. To overcome these challenges, the computational protein modeling suite Rosetta has been tailored to address various protein design tasks. Recently, statistical methods have been developed that identify correlated mutations between residues in a multiple sequence alignment of homologous proteins. These subtle inter-dependencies in the occupancy of residue positions throughout evolution are crucial for protein function, but we found that three current Rosetta design approaches fail to recover these co-evolutionary couplings. Thus, we developed the Rosetta method ResCue (residue-coupling enhanced) that leverages co-evolutionary information to favor sequences which recapitulate correlated mutations, as observed in nature. To assess the protocols via recapitulation designs, we compiled a benchmark of ten proteins each represented by two, structurally diverse states. We could demonstrate that ResCue designed sequences with an average sequence recovery rate of 70%, whereas three other protocols reached not more than 50%, on average. Our approach had higher recovery rates also for functionally important residues, which were studied in detail. This improvement has only a minor negative effect on the fitness of the designed sequences as assessed by Rosetta energy. In conclusion, our findings support the idea that informing protocols with co-evolutionary signals helps to design stable and native-like proteins that are compatible with the different conformational states required for a complex function.

In homologous proteins, functionally or structurally important residues are strongly conserved. Thus, the consideration of conservation signals during protein design protocols can help to create sequences that are more native-like. However, the number of conserved residues is small in many proteins and not all important residues can be captured by conservation analysis. Residues are forming networks whose composition is dictated by protein structure function and thus is visible through the co-evolutionary analysis. Nowadays, advanced methods allow us to deduce these networks from multiple sequence alignments. Thus, we have implemented the novel Rosetta method termed ‘ResCue’ that informs the design protocol with co-evolutionary signals. Recapitulation designs based on ten difficult benchmarks made clear that this protocol creates sequences that are more native-like than three other, state-of-the-art design protocols.

Decoding Protein-protein Interactions: An Overview

Drug discovery has focused on the paradigm "one drug, one target" for a long time. However, small molecules can act at multiple macromolecular targets, which serves as the basis for drug repurposing. In an effort to expand the target space, and given advances in X-ray crystallography, protein-protein interactions have become an emerging focus area of drug discovery enterprises. Proteins interact with other biomolecules and it is this intricate network of interactions that determines the behavior of the system and its biological processes. In this review, we briefly discuss networks in disease, followed by computational methods for protein-protein complex prediction. Computational methodologies and techniques employed towards objectives such as protein-protein docking, protein-protein interactions, and interface predictions are described extensively. Docking aims at producing a complex between proteins, while interface predictions identify a subset of residues on one protein that could interact with a partner, and protein-protein interaction sites address whether two proteins interact. In addition, approaches to predict hot spots and binding sites are presented along with a representative example of our internal project on the chemokine CXC receptor 3 B-isoform and predictive modeling with IP10 and PF4.