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

Recent progress and future challenges in structure-based protein-protein interaction prediction

Protein-protein interactions (PPIs) play a fundamental role in cellular processes, and understanding these interactions is crucial for advances in both basic biological science and biomedical applications. This review presents an overview of recent progress in computational methods for modeling protein complexes and predicting PPIs based on 3D structures, focusing on the transformative role of artificial intelligence-based approaches. We further discuss the expanding biomedical applications of PPI research, including the elucidation of disease mechanisms, drug discovery, and therapeutic design. Despite these advances, significant challenges remain in predicting host-pathogen interactions, interactions between intrinsically disordered regions, and interactions related to immune responses. These challenges are worthwhile for future explorations and represent the frontier of research in this field.

Transmembrane Homodimers Interface Identification: Predicting Interface Residues in Alpha-Helical Transmembrane Protein Homodimers Using Sequential and Structural Features

Most bitopic transmembrane proteins associate with one another through interface residues to form dimers, which facilitate or activate specific cellular functions. Therefore, accurately identifying interface residues in a given dimer is crucial for understanding its function and has been a challenging pursuit for many computational methods. These methods can be broadly categorized into two approaches: general-purpose ones for dimerization and specialized ones for interface residues. In this study, we develop a machine learning method that integrates both approaches by integrating sequential and structural features extracted from predicted structures and various domains. The results from cross-validation on a benchmark dataset show that our method, despite utilizing significantly fewer features, outperforms the state-of-the-art methods by more than three percentage points in performance, as measured by the F1 score. Furthermore, we evaluated the performance of the proposed model on a benchmark dataset as compared to the state-of-the-art multimeric structure predictors, including RoseTTAFold2, AlphaFold2Multimer, and PREDDIMER. The results show the superiority of the proposed model by outperforming all the other models, highlighting the effectiveness of integrating both structural and sequential features within the proposed framework.

EvoWeaver: large-scale prediction of gene functional associations from coevolutionary signals

The known universe of uncharacterized proteins is expanding far faster than our ability to annotate their functions through laboratory study. Computational annotation approaches rely on similarity to previously studied proteins, thereby ignoring unstudied proteins. Coevolutionary approaches hold promise for injecting new information into our knowledge of the protein universe by linking proteins through 'guilt-by-association'. However, existing coevolutionary algorithms have insufficient accuracy and scalability to connect the entire universe of proteins. We present EvoWeaver, a method that weaves together 12 signals of coevolution to quantify the degree of shared evolution between genes. EvoWeaver accurately identifies proteins involved in protein complexes or separate steps of a biochemical pathway. We show the merits of EvoWeaver by partly reconstructing known biochemical pathways without any prior knowledge other than that available from genomic sequences. Applying EvoWeaver to 1545 gene groups from 8564 genomes reveals missing connections in popular databases and potentially undiscovered links between proteins.

Emerging frontiers in protein structure prediction following the AlphaFold revolution

Models of protein structures enable molecular understanding of biological processes. Current protein structure prediction tools lie at the interface of biology, chemistry and computer science. Millions of protein structure models have been generated in a very short space of time through a revolution in protein structure prediction driven by deep learning, led by AlphaFold. This has provided a wealth of new structural information. Interpreting these predictions is critical to determining where and when this information is useful. But proteins are not static nor do they act alone, and structures of proteins interacting with other proteins and other biomolecules are critical to a complete understanding of their biological function at the molecular level. This review focuses on the application of state-of-the-art protein structure prediction to these advanced applications. We also suggest a set of guidelines for reporting AlphaFold predictions.

Identification of coevolving positions by ancestral reconstruction

Coevolution within proteins occurs when changes in one position affect the selective pressure in another position to preserve the protein structure or function. The identification of coevolving positions within proteins remains contentious, with most methods disregarding the phylogenetic information. Here, we present a time-efficient approach for detecting coevolving pairs, which is almost perfect in terms of precision and specificity. It is based on maximum parsimony-based ancestral reconstruction followed by the identification of pairs with a depletion on separate changes when compared to their number of concurrent changes. Our analysis of a previously characterised biological dataset shows that the coevolving pairs that we identified tend to be close in the protein sequence and structure, slightly less solvent exposed and have a higher mutation rate. We also show how the ancestral reconstruction can be used to detect favourable and unfavourable amino acid combinations. Altogether, we demonstrate how this approach is essential for identifying pairs of positions with weak covariation patterns.

Using residue interaction networks to understand protein function and evolution and to engineer new proteins

Residue interaction networks (RINs) provide graph-based representations of interaction networks within proteins, providing important insight into the factors driving protein structure, function, and stability relationships. There exists a wide range of tools with which to perform RIN analysis, taking into account different types of interactions, input (crystal structures, simulation trajectories, single proteins, or comparative analysis across proteins), as well as formats, including standalone software, web server, and a web application programming interface (API). In particular, the ability to perform comparative RIN analysis across protein families using "metaRINs" provides a valuable tool with which to dissect protein evolution. This, in turn, highlights hotspots to avoid (or target) for in vitro evolutionary studies, providing a powerful framework that can be exploited to engineer new proteins.

Prediction of Inter-Residue Multiple Distances and Exploration of Protein Multiple Conformations by Deep Learning

AlphaFold2 has achieved a major breakthrough in end-to-end prediction for static protein structures. However, protein conformational change is considered to be a key factor in protein biological function. Inter-residue multiple distances prediction is of great significance for research on protein multiple conformations exploration. In this study, we proposed an inter-residue multiple distances prediction method, DeepMDisPre, based on an improved network which integrates triangle update, axial attention and ResNet to predict multiple distances of residue pairs. We built a dataset which contains proteins with a single structure and proteins with multiple conformations to train the network. We tested DeepMDisPre on 114 proteins with multiple conformations. The results show that the inter-residue distance distribution predicted by DeepMDisPre tends to have multiple peaks for flexible residue pairs than for rigid residue pairs. On two cases of proteins with multiple conformations, we modeled the multiple conformations relatively accurately by using the predicted inter-residue multiple distances. In addition, we also tested the performance of DeepMDisPre on 279 proteins with a single structure. Experimental results demonstrate that the average contact accuracy of DeepMDisPre is higher than that of the comparative method. In terms of static protein modeling, the average TM-score of the 3D models built by DeepMDisPre is also improved compared with the comparative method.

Artificial Intelligence Learns Protein Prediction

From AlphaGO over StableDiffusion to ChatGPT, the recent decade of exponential advances in artificial intelligence (AI) has been altering life. In parallel, advances in computational biology are beginning to decode the language of life: AlphaFold2 leaped forward in protein structure prediction, and protein language models (pLMs) replaced expertise and evolutionary information from multiple sequence alignments with information learned from reoccurring patterns in databases of billions of proteins without experimental annotations other than the amino acid sequences. None of those tools could have been developed 10 years ago; all will increase the wealth of experimental data and speed up the cycle from idea to proof. AI is affecting molecular and medical biology at giant steps, and the most important might be the leap toward more powerful protein design.

Unveiling the evolution of policies for enhancing protein structure predictions: A comprehensive analysis

Predicting protein structure is both fascinating and formidable, playing a crucial role in structure-based drug discovery and unraveling diseases with elusive origins. The Critical Assessment of Protein Structure Prediction (CASP) serves as a biannual battleground where global scientists converge to untangle the intricate relationships within amino acid chains. Two primary methods, Template-Based Modeling (TBM) and Template-Free (TF) strategies, dominate protein structure prediction. The trend has shifted towards Template-Free predictions due to their broader sequence coverage with fewer templates. The predictive process can be broadly classified into contact map, binned-distance, and real-valued distance predictions, each with distinctive strengths and limitations manifested through tailored loss functions. We have also introduced revolutionary end-to-end, and all-atom diffusion-based techniques that have transformed protein structure predictions. Recent advancements in deep learning techniques have significantly improved prediction accuracy, although the effectiveness is contingent upon the quality of input features derived from natural bio-physiochemical attributes and Multiple Sequence Alignments (MSA). Hence, the generation of high-quality MSA data holds paramount importance in harnessing informative input features for enhanced prediction outcomes. Remarkable successes have been achieved in protein structure prediction accuracy, however not enough for what structural knowledge was intended to, which implies need for development in some other aspects of the predictions. In this regard, scientists have opened other frontiers for protein structural prediction. The utilization of subsampling in multiple sequence alignment (MSA) and protein language modeling appears to be particularly promising in enhancing the accuracy and efficiency of predictions, ultimately aiding in drug discovery efforts. The exploration of predicting protein complex structure also opens up exciting opportunities to deepen our knowledge of molecular interactions and design therapeutics that are more effective. In this article, we have discussed the vicissitudes that the scientists have gone through to improve prediction accuracy, and examined the effective policies in predicting from different aspects, including the construction of high quality MSA, providing informative input features, and progresses in deep learning approaches. We have also briefly touched upon transitioning from predicting single-chain protein structures to predicting protein complex structures. Our findings point towards promoting open research environments to support the objectives of protein structure prediction.

AlphaFold two years on: Validation and impact

Two years on from the initial release of AlphaFold, we have seen its widespread adoption as a structure prediction tool. Here, we discuss some of the latest work based on AlphaFold, with a particular focus on its use within the structural biology community. This encompasses use cases like speeding up structure determination itself, enabling new computational studies, and building new tools and workflows. We also look at the ongoing validation of AlphaFold, as its predictions continue to be compared against large numbers of experimental structures to further delineate the model's capabilities and limitations.

Machine learning in biological physics: From biomolecular prediction to design

Machine learning has been proposed as an alternative to theoretical modeling when dealing with complex problems in biological physics. However, in this perspective, we argue that a more successful approach is a proper combination of these two methodologies. We discuss how ideas coming from physical modeling neuronal processing led to early formulations of computational neural networks, e.g., Hopfield networks. We then show how modern learning approaches like Potts models, Boltzmann machines, and the transformer architecture are related to each other, specifically, through a shared energy representation. We summarize recent efforts to establish these connections and provide examples on how each of these formulations integrating physical modeling and machine learning have been successful in tackling recent problems in biomolecular structure, dynamics, function, evolution, and design. Instances include protein structure prediction; improvement in computational complexity and accuracy of molecular dynamics simulations; better inference of the effects of mutations in proteins leading to improved evolutionary modeling and finally how machine learning is revolutionizing protein engineering and design. Going beyond naturally existing protein sequences, a connection to protein design is discussed where synthetic sequences are able to fold to naturally occurring motifs driven by a model rooted in physical principles. We show that this model is "learnable" and propose its future use in the generation of unique sequences that can fold into a target structure.

Assessing the functional roles of coevolving PHD finger residues

Although in silico folding based on coevolving residue constraints in the deep-learning era has transformed protein structure prediction, the contributions of coevolving residues to protein folding, stability, and other functions in physical contexts remain to be clarified and experimentally validated. Herein, the PHD finger module, a well-known histone reader with distinct subtypes containing subtype-specific coevolving residues, was used as a model to experimentally assess the contributions of coevolving residues and to clarify their specific roles. The results of the assessment, including proteolysis and thermal unfolding of wildtype and mutant proteins, suggested that coevolving residues have varying contributions, despite their large in silico constraints. Residue positions with large constraints were found to contribute to stability in one subtype but not others. Computational sequence design and generative model-based energy estimates of individual structures were also implemented to complement the experimental assessment. Sequence design and energy estimates distinguish coevolving residues that contribute to folding from those that do not. The results of proteolytic analysis of mutations at positions contributing to folding were consistent with those suggested by sequence design and energy estimation. Thus, we report a comprehensive assessment of the contributions of coevolving residues, as well as a strategy based on a combination of approaches that should enable detailed understanding of the residue contributions in other large protein families.

Using Attention-UNet Models to Predict Protein Contact Maps

Proteins are essential to life, and understanding their intrinsic roles requires determining their structure. The field of proteomics has opened up new opportunities by applying deep learning algorithms to large databases of solved protein structures. With the availability of large data sets and advanced machine learning methods, the prediction of protein residue interactions has greatly improved. Protein contact maps provide empirical evidence of the interacting residue pairs within a protein sequence. Template-free protein structure prediction systems rely heavily on this information. This article proposes UNet-CON, an attention-integrated UNet architecture, trained to predict residue-residue contacts in protein sequences. With the predicted contacts being more accurate than state-of-the-art methods on the PDB25 test set, the model paves the way for the development of more powerful deep learning algorithms for predicting protein residue interactions.

ZEPPI: Proteome-scale sequence-based evaluation of protein-protein interaction models

We introduce ZEPPI (Z-score Evaluation of Protein-Protein Interfaces), a framework to evaluate structural models of a complex based on sequence coevolution and conservation involving residues in protein-protein interfaces. The ZEPPI score is calculated by comparing metrics for an interface to those obtained from randomly chosen residues. Since contacting residues are defined by the structural model, this obviates the need to account for indirect interactions. Further, although ZEPPI relies on species-paired multiple sequence alignments, its focus on interfacial residues allows it to leverage quite shallow alignments. ZEPPI can be implemented on a proteome-wide scale and is applied here to millions of structural models of dimeric complexes in the Escherichia coli and human interactomes found in the PrePPI database. PrePPI's scoring function is based primarily on the evaluation of protein-protein interfaces, and ZEPPI adds a new feature to this analysis through the incorporation of evolutionary information. ZEPPI performance is evaluated through applications to experimentally determined complexes and to decoys from the CASP-CAPRI experiment. As we discuss, the standard CAPRI scores used to evaluate docking models are based on model quality and not on the ability to give yes/no answers as to whether two proteins interact. ZEPPI is able to detect weak signals from PPI models that the CAPRI scores define as incorrect and, similarly, to identify potential PPIs defined as low confidence by the current PrePPI scoring function. A number of examples that illustrate how the combination of PrePPI and ZEPPI can yield functional hypotheses are provided.

A Perspective on Protein Structure Prediction Using Quantum Computers

Despite the recent advancements by deep learning methods such as AlphaFold2, in silico protein structure prediction remains a challenging problem in biomedical research. With the rapid evolution of quantum computing, it is natural to ask whether quantum computers can offer some meaningful benefits for approaching this problem. Yet, identifying specific problem instances amenable to quantum advantage and estimating the quantum resources required are equally challenging tasks. Here, we share our perspective on how to create a framework for systematically selecting protein structure prediction problems that are amenable for quantum advantage, and estimate quantum resources for such problems on a utility-scale quantum computer. As a proof-of-concept, we validate our problem selection framework by accurately predicting the structure of a catalytic loop of the Zika Virus NS3 Helicase, on quantum hardware.

Predicting hotspots for disease-causing single nucleotide variants using sequences-based coevolution, network analysis, and machine learning

To enable personalized medicine, it is important yet highly challenging to accurately predict disease-causing mutations in target proteins at high throughput. Previous computational methods have been developed using evolutionary information in combination with various biochemical and structural features of protein residues to discriminate neutral vs. deleterious mutations. However, the power of these methods is often limited because they either assume known protein structures or treat residues independently without fully considering their interactions. To address the above limitations, we build upon recent progress in machine learning, network analysis, and protein language models, and develop a sequences-based variant site prediction workflow based on the protein residue contact networks: 1. We employ and integrate various methods of building protein residue networks using state-of-the-art coevolution analysis tools (RaptorX, DeepMetaPSICOV, and SPOT-Contact) powered by deep learning. 2. We use machine learning algorithms (Random Forest, Gradient Boosting, and Extreme Gradient Boosting) to optimally combine 20 network centrality scores to jointly predict key residues as hot spots for disease mutations. 3. Using a dataset of 107 proteins rich in disease mutations, we rigorously evaluate the network scores individually and collectively (via machine learning). This work supports a promising strategy of combining an ensemble of network scores based on different coevolution analysis methods (and optionally predictive scores from other methods) via machine learning to predict hotspot sites of disease mutations, which will inform downstream applications of disease diagnosis and targeted drug design.

Can Protein Structure Prediction Methods Capture Alternative Conformations of Membrane Transporters?

Understanding the conformational dynamics of proteins, such as the inward-facing (IF) and outward-facing (OF) transition observed in transporters, is vital for elucidating their functional mechanisms. Despite significant advances in protein structure prediction (PSP) over the past three decades, most efforts have been focused on single-state prediction, leaving multistate or alternative conformation prediction (ACP) relatively unexplored. This discrepancy has led to the development of highly accurate PSP methods such as AlphaFold, yet their capabilities for ACP remain limited. To investigate the performance of current PSP methods in ACP, we curated a data set, named IOMemP, consisting of 32 experimentally determined high-resolution IF and OF structures of 16 membrane proteins with substantial conformational changes. We benchmarked 12 representative PSP methods, along with two recent multistate methods based on AlphaFold, against this data set. Our findings reveal a remarkably consistent preference for specific states across various PSP methods. We elucidated how coevolution information in MSAs influences state preference. Moreover, we showed that AlphaFold, when excluding coevolution information, estimated similar energies between the experimental IF and OF conformations, indicating that the energy model learned by AlphaFold is not biased toward any particular state. Our IOMemP data set and benchmark results are anticipated to advance the development of robust ACP methods.

An interactive visualization tool for educational outreach in protein contact map overlap analysis

Recent advancements in contact map-based protein three-dimensional (3D) structure prediction have been driven by the evolution of deep learning algorithms. However, the gap in accessible software tools for novices in this domain remains a significant challenge. This study introduces GoFold, a novel, standalone graphical user interface (GUI) designed for beginners to perform contact map overlap (CMO) problems for better template selection. Unlike existing tools that cater more to research needs or assume foundational knowledge, GoFold offers an intuitive, user-friendly platform with comprehensive tutorials. It stands out in its ability to visually represent the CMO problem, allowing users to input proteins in various formats and explore the CMO problem. The educational value of GoFold is demonstrated through benchmarking against the state-of-the-art contact map overlap method, map_align, using two datasets: PSICOV and CAMEO. GoFold exhibits superior performance in terms of TM-score and Z-score metrics across diverse qualities of contact maps and target difficulties. Notably, GoFold runs efficiently on personal computers without any third-party dependencies, thereby making it accessible to the general public for promoting citizen science. The tool is freely available for download for macOS, Linux, and Windows.

Enhancing coevolutionary signals in protein-protein interaction prediction through clade-wise alignment integration

Protein-protein interactions (PPIs) play essential roles in most biological processes. The binding interfaces between interacting proteins impose evolutionary constraints that have successfully been employed to predict PPIs from multiple sequence alignments (MSAs). To construct MSAs, critical choices have to be made: how to ensure the reliable identification of orthologs, and how to optimally balance the need for large alignments versus sufficient alignment quality. Here, we propose a divide-and-conquer strategy for MSA generation: instead of building a single, large alignment for each protein, multiple distinct alignments are constructed under distinct clades in the tree of life. Coevolutionary signals are searched separately within these clades, and are only subsequently integrated using machine learning techniques. We find that this strategy markedly improves overall prediction performance, concomitant with better alignment quality. Using the popular DCA algorithm to systematically search pairs of such alignments, a genome-wide all-against-all interaction scan in a bacterial genome is demonstrated. Given the recent successes of AlphaFold in predicting direct PPIs at atomic detail, a discover-and-refine approach is proposed: our method could provide a fast and accurate strategy for pre-screening the entire genome, submitting to AlphaFold only promising interaction candidates-thus reducing false positives as well as computation time.

Key interaction networks: Identifying evolutionarily conserved non-covalent interaction networks across protein families

Protein structure (and thus function) is dictated by non-covalent interaction networks. These can be highly evolutionarily conserved across protein families, the members of which can diverge in sequence and evolutionary history. Here we present KIN, a tool to identify and analyze conserved non-covalent interaction networks across evolutionarily related groups of proteins. KIN is available for download under a GNU General Public License, version 2, from https://www.github.com/kamerlinlab/KIN. KIN can operate on experimentally determined structures, predicted structures, or molecular dynamics trajectories, providing insight into both conserved and missing interactions across evolutionarily related proteins. This provides useful insight both into protein evolution, as well as a tool that can be exploited for protein engineering efforts. As a showcase system, we demonstrate applications of this tool to understanding the evolutionary-relevant conserved interaction networks across the class A β-lactamases.

Drug-target affinity prediction with extended graph learning-convolutional networks

BACKGROUND

High-performance computing plays a pivotal role in computer-aided drug design, a field that holds significant promise in pharmaceutical research. The prediction of drug-target affinity (DTA) is a crucial stage in this process, potentially accelerating drug development through rapid and extensive preliminary compound screening, while also minimizing resource utilization and costs. Recently, the incorporation of deep learning into DTA prediction and the enhancement of its accuracy have emerged as key areas of interest in the research community. Drugs and targets can be characterized through various methods, including structure-based, sequence-based, and graph-based representations. Despite the progress in structure and sequence-based techniques, they tend to provide limited feature information. Conversely, graph-based approaches have risen to prominence, attracting considerable attention for their comprehensive data representation capabilities. Recent studies have focused on constructing protein and drug molecular graphs using sequences and SMILES, subsequently deriving representations through graph neural networks. However, these graph-based approaches are limited by the use of a fixed adjacent matrix of protein and drug molecular graphs for graph convolution. This limitation restricts the learning of comprehensive feature representations from intricate compound and protein structures, consequently impeding the full potential of graph-based feature representation in DTA prediction. This, in turn, significantly impacts the models' generalization capabilities in the complex realm of drug discovery.

RESULTS

To tackle these challenges, we introduce GLCN-DTA, a model specifically designed for proficiency in DTA tasks. GLCN-DTA innovatively integrates a graph learning module into the existing graph architecture. This module is designed to learn a soft adjacent matrix, which effectively and efficiently refines the contextual structure of protein and drug molecular graphs. This advancement allows for learning richer structural information from protein and drug molecular graphs via graph convolution, specifically tailored for DTA tasks, compared to the conventional fixed adjacent matrix approach. A series of experiments have been conducted to validate the efficacy of the proposed GLCN-DTA method across diverse scenarios. The results demonstrate that GLCN-DTA possesses advantages in terms of robustness and high accuracy.

CONCLUSIONS

The proposed GLCN-DTA model enhances DTA prediction performance by introducing a novel framework that synergizes graph learning operations with graph convolution operations, thereby achieving richer representations. GLCN-DTA does not distinguish between different protein classifications, including structurally ordered and intrinsically disordered proteins, focusing instead on improving feature representation. Therefore, its applicability scope may be more effective in scenarios involving structurally ordered proteins, while potentially being limited in contexts with intrinsically disordered proteins.

A comparative study of protein structure prediction tools for challenging targets: Snake venom toxins

Protein structure determination is a critical aspect of biological research, enabling us to understand protein function and potential applications. Recent advances in deep learning and artificial intelligence have led to the development of several protein structure prediction tools, such as AlphaFold2 and ColabFold. However, their performance has primarily been evaluated on well-characterised proteins and their ability to predict sturtctures of proteins lacking experimental structures, such as many snake venom toxins, has been less scrutinised. In this study, we evaluated three modelling tools on their prediction of over 1000 snake venom toxin structures for which no experimental structures exist. Our findings show that AlphaFold2 (AF2) performed the best across all assessed parameters. We also observed that ColabFold (CF) only scored slightly worse than AF2, while being computationally less intensive. All tools struggled with regions of intrinsic disorder, such as loops and propeptide regions, and performed well in predicting the structure of functional domains. Overall, our study highlights the importance of exercising caution when working with proteins with no experimental structures available, particularly those that are large and contain flexible regions. Nonetheless, leveraging computational structure prediction tools can provide valuable insights into the modelling of protein interactions with different targets and reveal potential binding sites, active sites, and conformational changes, as well as into the design of potential molecular binders for reagent, diagnostic, or therapeutic purposes.

AttCON: With better MSAs and attention mechanism for accurate protein contact map prediction

Protein contact map prediction is a critical and vital step in protein structure prediction, and its accuracy is highly contingent upon the feature representations of protein sequence information and the efficacy of deep learning models. In this paper, we propose an algorithm, DeepMSA+, to generate protein multiple sequence alignments (MSAs) and to construct feature representations based on co-evolutionary information and sequence information derived from MSAs. We also propose an improved deep learning model, AttCON, for training input features to predict protein contact map. The model incorporates an attention module, and by comparing different attention modules, we find a parameter-free attention module suitable for contact map prediction. Additionally, we use the Focal Loss function to better address the data imbalance issue in protein contact map. We also developed a weighted evaluation index (W score) for model evaluation, which takes into account a wide range of metrics. W score is comprehensive in its scope, with a particular focus on the precision of predictions for medium-range and long-range contacts. Experimental results show that AttCON achieves good precision results on datasets from CASP11 to CASP15. Compared to some state-of-the-art methods, it achieves an average improvement of over 5% in both medium-range and long-range predictions, and W score is improved by an average of 2 points.

Review of Computational Methods and Database Sources for Predicting the Effects of Coding Frameshift Small Insertion and Deletion Variations

Genetic variations (including substitutions, insertions, and deletions) exert a profound influence on DNA sequences. These variations are systematically classified as synonymous, nonsynonymous, and nonsense, each manifesting distinct effects on proteins. The implementation of high-throughput sequencing has significantly augmented our comprehension of the intricate interplay between gene variations and protein structure and function, as well as their ramifications in the context of diseases. Frameshift variations, particularly small insertions and deletions (indels), disrupt protein coding and are instrumental in disease pathogenesis. This review presents a succinct review of computational methods, databases, current challenges, and future directions in predicting the consequences of coding frameshift small indels variations. We analyzed the predictive efficacy, reliability, and utilization of computational methods and variant account, reliability, and utilization of database. Besides, we also compared the prediction methodologies on GOF/LOF pathogenic variation data. Addressing the challenges pertaining to prediction accuracy and cross-species generalizability, nascent technologies such as AI and deep learning harbor immense potential to enhance predictive capabilities. The importance of interdisciplinary research and collaboration cannot be overstated for devising effective diagnosis, treatment, and prevention strategies concerning diseases associated with coding frameshift indels variations.

ConPep: Prediction of peptide contact maps with pre-trained biological language model and multi-view feature extracting strategy

The accurate prediction of peptide contact maps remains a challenging task due to the difficulty in obtaining the interactive information between residues on short sequences. To address this challenge, we propose ConPep, a deep learning framework designed for predicting the contact map of peptides based on sequences only. To sufficiently incorporate the sequential semantic information between residues in peptide sequences, we use a pre-trained biological language model and transfer prior knowledge from large scale databases. Additionally, to extract and integrate sequential local information and residue-based global correlations, our model incorporates Bidirectional Gated Recurrent Unit and attention mechanisms. They can obtain multi-view features and thus enhance the accuracy and robustness of our prediction. Comparative results on independent tests demonstrate that our proposed method significantly outperforms state-of-the-art methods even with short peptides. Notably, our method exhibits superior performance at the sequence level, suggesting the robust ability of our model compared with the multiple sequence alignment (MSA) analysis-based methods. We expect it can be meaningful research for facilitating the wide use of our method.

PepCNN deep learning tool for predicting peptide binding residues in proteins using sequence, structural, and language model features

Protein-peptide interactions play a crucial role in various cellular processes and are implicated in abnormal cellular behaviors leading to diseases such as cancer. Therefore, understanding these interactions is vital for both functional genomics and drug discovery efforts. Despite a significant increase in the availability of protein-peptide complexes, experimental methods for studying these interactions remain laborious, time-consuming, and expensive. Computational methods offer a complementary approach but often fall short in terms of prediction accuracy. To address these challenges, we introduce PepCNN, a deep learning-based prediction model that incorporates structural and sequence-based information from primary protein sequences. By utilizing a combination of half-sphere exposure, position specific scoring matrices from multiple-sequence alignment tool, and embedding from a pre-trained protein language model, PepCNN outperforms state-of-the-art methods in terms of specificity, precision, and AUC. The PepCNN software and datasets are publicly available at https://github.com/abelavit/PepCNN.git .

E2EDA: Protein Domain Assembly Based on End-to-End Deep Learning

With the development of deep learning, almost all single-domain proteins can be predicted at experimental resolution. However, the structure prediction of multi-domain proteins remains a challenge. Achieving end-to-end protein domain assembly and further improving the accuracy of the full-chain modeling by accurately predicting inter-domain orientation while improving the assembly efficiency will provide significant insights into structure-based drug discovery. In this work, we propose an End-to-End Domain Assembly method based on deep learning, named E2EDA. We first develop RMNet, an EfficientNetV2-based deep learning model that fuses multiple features using an attention mechanism to predict inter-domain rigid motion. Then, the predicted rigid motions are transformed into inter-domain spatial transformations to directly assemble the full-chain model. Finally, the scoring strategy RMscore is designed to select the best model from multiple assembled models. The experimental results show that the average TM-score of the model assembled by E2EDA on the benchmark set (282) is 0.827, which is better than those of other domain assembly methods SADA (0.792) and DEMO (0.730). Meanwhile, on our constructed multi-domain data set from AlphaFold DB, the model reassembled by E2EDA is 7.0% higher in TM-score compared to the full-chain model predicted by AlphaFold2, indicating that E2EDA can capture more accurate inter-domain orientations to improve the quality of the model predicted by AlphaFold2. Furthermore, compared to SADA and AlphaFold2, E2EDA reduced the average runtime on the benchmark by 64.7% and 19.2%, respectively, indicating that E2EDA can significantly improve assembly efficiency through an end-to-end approach. The online server is available at http://zhanglab-bioinf.com/E2EDA.

New alignment method for remote protein sequences by the direct use of pairwise sequence correlations and substitutions

Understanding protein sequences and how they relate to the functions of proteins is extremely important. One of the most basic operations in bioinformatics is sequence alignment and usually the first things learned from these are which positions are the most conserved and often these are critical parts of the structure, such as enzyme active site residues. In addition, the contact pairs in a protein usually correspond closely to the correlations between residue positions in the multiple sequence alignment, and these usually change in a systematic and coordinated way, if one position changes then the other member of the pair also changes to compensate. In the present work, these correlated pairs are taken as anchor points for a new type of sequence alignment. The main advantage of the method here is its combining the remote homolog detection from our method PROST with pairwise sequence substitutions in the rigorous method from Kleinjung et al. We show a few examples of some resulting sequence alignments, and how they can lead to improvements in alignments for function, even for a disordered protein.

Protein Structure Prediction: Challenges, Advances, and the Shift of Research Paradigms

Protein structure prediction is an interdisciplinary research topic that has attracted researchers from multiple fields, including biochemistry, medicine, physics, mathematics, and computer science. These researchers adopt various research paradigms to attack the same structure prediction problem: biochemists and physicists attempt to reveal the principles governing protein folding; mathematicians, especially statisticians, usually start from assuming a probability distribution of protein structures given a target sequence and then find the most likely structure, while computer scientists formulate protein structure prediction as an optimization problem - finding the structural conformation with the lowest energy or minimizing the difference between predicted structure and native structure. These research paradigms fall into the two statistical modeling cultures proposed by Leo Breiman, namely, data modeling and algorithmic modeling. Recently, we have also witnessed the great success of deep learning in protein structure prediction. In this review, we present a survey of the efforts for protein structure prediction. We compare the research paradigms adopted by researchers from different fields, with an emphasis on the shift of research paradigms in the era of deep learning. In short, the algorithmic modeling techniques, especially deep neural networks, have considerably improved the accuracy of protein structure prediction; however, theories interpreting the neural networks and knowledge on protein folding are still highly desired.

DeepBSRPred: deep learning-based binding site residue prediction for proteins

MOTIVATION

Proteins-protein interactions (PPIs) are important to govern several cellular activities. Amino acid residues, which are located at the interface are known as the binding sites and the information about binding sites helps to understand the binding affinities and functions of protein-protein complexes.

RESULTS

We have developed a deep neural network-based method, DeepBSRPred, for predicting the binding sites using protein sequence information and predicted structures from AlphaFold2. Specific sequence and structure-based features include position-specific scoring matrix (PSSM), solvent accessible surface area, conservation score and amino acid properties, and residue depth, respectively. Our method predicted the binding sites with an average F1 score of 0.73 in a dataset of 1236 proteins. Further, we compared the performance with other existing methods in the literature using four benchmark datasets and our method outperformed those methods.

AVAILABILITY AND IMPLEMENTATION

The DeepBSRPred web server can be found at https://web.iitm.ac.in/bioinfo2/deepbsrpred/index.html , along with all datasets used in this study. The trained models, the DeepBSRPred standalone source code, and the feature computation pipeline are freely available at https://web.iitm.ac.in/bioinfo2/deepbsrpred/download.html .

ETLD: an encoder-transformation layer-decoder architecture for protein contact and mutation effects prediction

The latent features extracted from the multiple sequence alignments (MSAs) of homologous protein families are useful for identifying residue-residue contacts, predicting mutation effects, shaping protein evolution, etc. Over the past three decades, a growing body of supervised and unsupervised machine learning methods have been applied to this field, yielding fruitful results. Here, we propose a novel self-supervised model, called encoder-transformation layer-decoder (ETLD) architecture, capable of capturing protein sequence latent features directly from MSAs. Compared to the typical autoencoder model, ETLD introduces a transformation layer with the ability to learn inter-site couplings, which can be used to parse out the two-dimensional residue-residue contacts map after a simple mathematical derivation or an additional supervised neural network. ETLD retains the process of encoding and decoding sequences, and the predicted probabilities of amino acids at each site can be further used to construct the mutation landscapes for mutation effects prediction, outperforming advanced models such as GEMME, DeepSequence and EVmutation in general. Overall, ETLD is a highly interpretable unsupervised model with great potential for improvement and can be further combined with supervised methods for more extensive and accurate predictions.

Evolutionary selection of proteins with two folds

Although most globular proteins fold into a single stable structure, an increasing number have been shown to remodel their secondary and tertiary structures in response to cellular stimuli. State-of-the-art algorithms predict that these fold-switching proteins adopt only one stable structure, missing their functionally critical alternative folds. Why these algorithms predict a single fold is unclear, but all of them infer protein structure from coevolved amino acid pairs. Here, we hypothesize that coevolutionary signatures are being missed. Suspecting that single-fold variants could be masking these signatures, we developed an approach, called Alternative Contact Enhancement (ACE), to search both highly diverse protein superfamilies-composed of single-fold and fold-switching variants-and protein subfamilies with more fold-switching variants. ACE successfully revealed coevolution of amino acid pairs uniquely corresponding to both conformations of 56/56 fold-switching proteins from distinct families. Then, we used ACE-derived contacts to (1) predict two experimentally consistent conformations of a candidate protein with unsolved structure and (2) develop a blind prediction pipeline for fold-switching proteins. The discovery of widespread dual-fold coevolution indicates that fold-switching sequences have been preserved by natural selection, implying that their functionalities provide evolutionary advantage and paving the way for predictions of diverse protein structures from single sequences.

Discriminating physiological from non-physiological interfaces in structures of protein complexes: A community-wide study

Reliably scoring and ranking candidate models of protein complexes and assigning their oligomeric state from the structure of the crystal lattice represent outstanding challenges. A community-wide effort was launched to tackle these challenges. The latest resources on protein complexes and interfaces were exploited to derive a benchmark dataset consisting of 1677 homodimer protein crystal structures, including a balanced mix of physiological and non-physiological complexes. The non-physiological complexes in the benchmark were selected to bury a similar or larger interface area than their physiological counterparts, making it more difficult for scoring functions to differentiate between them. Next, 252 functions for scoring protein-protein interfaces previously developed by 13 groups were collected and evaluated for their ability to discriminate between physiological and non-physiological complexes. A simple consensus score generated using the best performing score of each of the 13 groups, and a cross-validated Random Forest (RF) classifier were created. Both approaches showed excellent performance, with an area under the Receiver Operating Characteristic (ROC) curve of 0.93 and 0.94, respectively, outperforming individual scores developed by different groups. Additionally, AlphaFold2 engines recalled the physiological dimers with significantly higher accuracy than the non-physiological set, lending support to the reliability of our benchmark dataset annotations. Optimizing the combined power of interface scoring functions and evaluating it on challenging benchmark datasets appears to be a promising strategy.

Two sequence- and two structure-based ML models have learned different aspects of protein biochemistry

Deep learning models are seeing increased use as methods to predict mutational effects or allowed mutations in proteins. The models commonly used for these purposes include large language models (LLMs) and 3D Convolutional Neural Networks (CNNs). These two model types have very different architectures and are commonly trained on different representations of proteins. LLMs make use of the transformer architecture and are trained purely on protein sequences whereas 3D CNNs are trained on voxelized representations of local protein structure. While comparable overall prediction accuracies have been reported for both types of models, it is not known to what extent these models make comparable specific predictions and/or generalize protein biochemistry in similar ways. Here, we perform a systematic comparison of two LLMs and two structure-based models (CNNs) and show that the different model types have distinct strengths and weaknesses. The overall prediction accuracies are largely uncorrelated between the sequence- and structure-based models. Overall, the two structure-based models are better at predicting buried aliphatic and hydrophobic residues whereas the two LLMs are better at predicting solvent-exposed polar and charged amino acids. Finally, we find that a combined model that takes the individual model predictions as input can leverage these individual model strengths and results in significantly improved overall prediction accuracy.

Deploying synthetic coevolution and machine learning to engineer protein-protein interactions

Fine-tuning of protein-protein interactions occurs naturally through coevolution, but this process is difficult to recapitulate in the laboratory. We describe a platform for synthetic protein-protein coevolution that can isolate matched pairs of interacting muteins from complex libraries. This large dataset of coevolved complexes drove a systems-level analysis of molecular recognition between Z domain-affibody pairs spanning a wide range of structures, affinities, cross-reactivities, and orthogonalities, and captured a broad spectrum of coevolutionary networks. Furthermore, we harnessed pretrained protein language models to expand, in silico, the amino acid diversity of our coevolution screen, predicting remodeled interfaces beyond the reach of the experimental library. The integration of these approaches provides a means of simulating protein coevolution and generating protein complexes with diverse molecular recognition properties for biotechnology and synthetic biology.

Improved structure-related prediction for insufficient homologous proteins using MSA enhancement and pre-trained language model

In recent years, protein structure problems have become a hotspot for understanding protein folding and function mechanisms. It has been observed that most of the protein structure works rely on and benefit from co-evolutionary information obtained by multiple sequence alignment (MSA). As an example, AlphaFold2 (AF2) is a typical MSA-based protein structure tool which is famous for its high accuracy. As a consequence, these MSA-based methods are limited by the quality of the MSAs. Especially for orphan proteins that have no homologous sequence, AlphaFold2 performs unsatisfactorily as MSA depth decreases, which may pose a barrier to its widespread application in protein mutation and design problems in which there are no rich homologous sequences and rapid prediction is needed. In this paper, we constructed two standard datasets for orphan and de novo proteins which have insufficient/none homology information, called Orphan62 and Design204, respectively, to fairly evaluate the performance of the various methods in this case. Then, depending on whether or not utilizing scarce MSA information, we summarized two approaches, MSA-enhanced and MSA-free methods, to effectively solve the issue without sufficient MSAs. MSA-enhanced model aims to improve poor MSA quality from the data source by knowledge distillation and generation models. MSA-free model directly learns the relationship between residues on enormous protein sequences from pre-trained models, bypassing the step of extracting the residue pair representation from MSA. Next, we evaluated the performance of four MSA-free methods (trRosettaX-Single, TRFold, ESMFold and ProtT5) and MSA-enhanced (Bagging MSA) method compared with a traditional MSA-based method AlphaFold2, in two protein structure-related prediction tasks, respectively. Comparison analyses show that trRosettaX-Single and ESMFold which belong to MSA-free method can achieve fast prediction ($\sim\! 40$s) and comparable performance compared with AF2 in tertiary structure prediction, especially for short peptides, $\alpha $-helical segments and targets with few homologous sequences. Bagging MSA utilizing MSA enhancement improves the accuracy of our trained base model which is an MSA-based method when poor homology information exists in secondary structure prediction. Our study provides biologists an insight of how to select rapid and appropriate prediction tools for enzyme engineering and peptide drug development.

CONTACT

guofei@csu.edu.cn, jj.tang@siat.ac.cn.

Thirteen dubious ways to detect conserved structural RNAs

Covariation induced by compensatory base substitutions in RNA alignments is a great way to deduce conserved RNA structure, in principle. In practice, success depends on many factors, importantly the quality and depth of the alignment and the choice of covariation statistic. Measuring covariation between pairs of aligned positions is easy. However, using covariation to infer evolutionarily conserved RNA structure is complicated by other extraneous sources of covariation such as that resulting from homologous sequences having evolved from a common ancestor. In order to provide evidence of evolutionarily conserved RNA structure, a method to distinguish covariation due to sources other than RNA structure is necessary. Moreover, there are several sorts of artifactually generated covariation signals that can further confound the analysis. Additionally, some covariation signal is difficult to detect due to incomplete comparative data. Here, we investigate and critically discuss the practice of inferring conserved RNA structure by comparative sequence analysis. We provide new methods on how to approach and decide which of the numerous long non-coding RNAs (lncRNAs) have biologically relevant structures.

Peptides of a Feather: How Computation Is Taking Peptide Therapeutics under Its Wing

Leveraging computation in the development of peptide therapeutics has garnered increasing recognition as a valuable tool to generate novel therapeutics for disease-related targets. To this end, computation has transformed the field of peptide design through identifying novel therapeutics that exhibit enhanced pharmacokinetic properties and reduced toxicity. The process of in-silico peptide design involves the application of molecular docking, molecular dynamics simulations, and machine learning algorithms. Three primary approaches for peptide therapeutic design including structural-based, protein mimicry, and short motif design have been predominantly adopted. Despite the ongoing progress made in this field, there are still significant challenges pertaining to peptide design including: enhancing the accuracy of computational methods; improving the success rate of preclinical and clinical trials; and developing better strategies to predict pharmacokinetics and toxicity. In this review, we discuss past and present research pertaining to the design and development of in-silico peptide therapeutics in addition to highlighting the potential of computation and artificial intelligence in the future of disease therapeutics.

Knowledge-guided data mining on the standardized architecture of NRPS: Subtypes, novel motifs, and sequence entanglements

Non-ribosomal peptide synthetase (NRPS) is a diverse family of biosynthetic enzymes for the assembly of bioactive peptides. Despite advances in microbial sequencing, the lack of a consistent standard for annotating NRPS domains and modules has made data-driven discoveries challenging. To address this, we introduced a standardized architecture for NRPS, by using known conserved motifs to partition typical domains. This motif-and-intermotif standardization allowed for systematic evaluations of sequence properties from a large number of NRPS pathways, resulting in the most comprehensive cross-kingdom C domain subtype classifications to date, as well as the discovery and experimental validation of novel conserved motifs with functional significance. Furthermore, our coevolution analysis revealed important barriers associated with re-engineering NRPSs and uncovered the entanglement between phylogeny and substrate specificity in NRPS sequences. Our findings provide a comprehensive and statistically insightful analysis of NRPS sequences, opening avenues for future data-driven discoveries.

Co-evolution-based prediction of metal-binding sites in proteomes by machine learning

Metal ions have various important biological roles in proteins, including structural maintenance, molecular recognition and catalysis. Previous methods of predicting metal-binding sites in proteomes were based on either sequence or structural motifs. Here we developed a co-evolution-based pipeline named 'MetalNet' to systematically predict metal-binding sites in proteomes. We applied MetalNet to proteomes of four representative prokaryotic species and predicted 4,849 potential metalloproteins, which substantially expands the currently annotated metalloproteomes. We biochemically and structurally validated previously unannotated metal-binding sites in several proteins, including apo-citrate lyase phosphoribosyl-dephospho-CoA transferase citX, an Escherichia coli enzyme lacking structural or sequence homology to any known metalloprotein (Protein Data Bank (PDB) codes: 7DCM and 7DCN ). MetalNet also successfully recapitulated all known zinc-binding sites from the human spliceosome complex. The pipeline of MetalNet provides a unique and enabling tool for interrogating the hidden metalloproteome and studying metal biology.

Latent generative landscapes as maps of functional diversity in protein sequence space

Variational autoencoders are unsupervised learning models with generative capabilities, when applied to protein data, they classify sequences by phylogeny and generate de novo sequences which preserve statistical properties of protein composition. While previous studies focus on clustering and generative features, here, we evaluate the underlying latent manifold in which sequence information is embedded. To investigate properties of the latent manifold, we utilize direct coupling analysis and a Potts Hamiltonian model to construct a latent generative landscape. We showcase how this landscape captures phylogenetic groupings, functional and fitness properties of several systems including Globins, β-lactamases, ion channels, and transcription factors. We provide support on how the landscape helps us understand the effects of sequence variability observed in experimental data and provides insights on directed and natural protein evolution. We propose that combining generative properties and functional predictive power of variational autoencoders and coevolutionary analysis could be beneficial in applications for protein engineering and design.

Statistical genetics in and out of quasi-linkage equilibrium

This review is about statistical genetics, an interdisciplinary topic between statistical physics and population biology. The focus is on the phase ofquasi-linkage equilibrium(QLE). Our goals here are to clarify under which conditions the QLE phase can be expected to hold in population biology and how the stability of the QLE phase is lost. The QLE state, which has many similarities to a thermal equilibrium state in statistical mechanics, was discovered by M Kimura for a two-locus two-allele model, and was extended and generalized to the global genome scale byNeher&Shraiman (2011). What we will refer to as the Kimura-Neher-Shraiman theory describes a population evolving due to the mutations, recombination, natural selection and possibly genetic drift. A QLE phase exists at sufficiently high recombination rate (r) and/or mutation ratesµwith respect to selection strength. We show how in QLE it is possible to infer the epistatic parameters of the fitness function from the knowledge of the (dynamical) distribution of genotypes in a population. We further consider the breakdown of the QLE regime for high enough selection strength. We review recent results for the selection-mutation and selection-recombination dynamics. Finally, we identify and characterize a new phase which we call the non-random coexistence where variability persists in the population without either fixating or disappearing.

Novel machine learning approaches revolutionize protein knowledge

Breakthrough methods in machine learning (ML), protein structure prediction, and novel ultrafast structural aligners are revolutionizing structural biology. Obtaining accurate models of proteins and annotating their functions on a large scale is no longer limited by time and resources. The most recent method to be top ranked by the Critical Assessment of Structure Prediction (CASP) assessment, AlphaFold 2 (AF2), is capable of building structural models with an accuracy comparable to that of experimental structures. Annotations of 3D models are keeping pace with the deposition of the structures due to advancements in protein language models (pLMs) and structural aligners that help validate these transferred annotations. In this review we describe how recent developments in ML for protein science are making large-scale structural bioinformatics available to the general scientific community.

Statistical Learning of Protein Elastic Network from Positional Covariance Matrix

Positional fluctuation and covariance during protein dynamics are key observables for understanding the molecular origin of biological functions. A frequently employed potential energy function for describing protein structural variation at the coarse-gained level is elastic network model (ENM). A long-standing issue in biomolecular simulation is thus the parametrization of ENM spring constants from the components of positional covariance matrix (PCM). Based on sensitivity analysis of PCM, the direct-coupling statistics of each spring, which is a specific combination of position fluctuation and covariance, is found to exhibit prominent signal of parameter dependence. This finding provides the basis for devising the objective function and the scheme of running through the effective one-dimensional optimization of every spring by self-consistent iteration. Formal derivation of the positional covariance statistical learning (PCSL) method also motivates the necessary data regularization for stable calculations. Robust convergence of PCSL is achieved in taking an all-atom molecular dynamics trajectory or an ensemble of homologous structures as input data. The PCSL framework can also be generalized with mixed objective functions to capture specific property such as the residue flexibility profile. Such physical chemistry-based statistical learning thus provides a useful platform for integrating the mechanical information encoded in various experimental or computational data.

Chasing long-range evolutionary couplings in the AlphaFold era

Coevolution between protein residues is normally interpreted as direct contact. However, the evolutionary record of a protein sequence contains rich information that may include long-range functional couplings, couplings that report on homo-oligomeric states or even conformational changes. Due to the complexity of the sequence space and the lack of structural information on various members of a protein family, it has been difficult to effectively mine the additional information encoded in a multiple sequence alignment (MSA). Here, taking advantage of the recent release of the AlphaFold (AF) database we attempt to identify coevolutionary couplings that cannot be explained simply by spatial proximity. We propose a simple computational method that performs direct coupling analysis on a MSA and searches for couplings that are not satisfied in any of the AF models of members of the identified protein family. Application of this method on 2012 protein families suggests that ~12% of the total identified coevolving residue pairs are spatially distant and more likely to be disordered than their contacting counterparts. We expect that this analysis will help improve the quality of coevolutionary distance restraints used for structure determination and will be useful in identifying potentially functional/allosteric cross-talk between distant residues.

Machine learning in computational modelling of membrane protein sequences and structures: From methodologies to applications

Membrane proteins mediate a wide spectrum of biological processes, such as signal transduction and cell communication. Due to the arduous and costly nature inherent to the experimental process, membrane proteins have long been devoid of well-resolved atomic-level tertiary structures and, consequently, the understanding of their functional roles underlying a multitude of life activities has been hampered. Currently, computational tools dedicated to furthering the structure-function understanding are primarily focused on utilizing intelligent algorithms to address a variety of site-wise prediction problems (e.g., topology and interaction sites), but are scattered across different computing sources. Moreover, the recent advent of deep learning techniques has immensely expedited the development of computational tools for membrane protein-related prediction problems. Given the growing number of applications optimized particularly by manifold deep neural networks, we herein provide a review on the current status of computational strategies mainly in membrane protein type classification, topology identification, interaction site detection, and pathogenic effect prediction. Meanwhile, we provide an overview of how the entire prediction process proceeds, including database collection, data pre-processing, feature extraction, and method selection. This review is expected to be useful for developing more extendable computational tools specific to membrane proteins.

Evolutionary selection of proteins with two folds

Although most globular proteins fold into a single stable structure 1 , an increasing number have been shown to remodel their secondary and tertiary structures in response to cellular stimuli 2 . State-of-the-art algorithms 3-5 predict that these fold-switching proteins assume only one stable structure 6,7 , missing their functionally critical alternative folds. Why these algorithms predict a single fold is unclear, but all of them infer protein structure from coevolved amino acid pairs. Here, we hypothesize that coevolutionary signatures are being missed. Suspecting that over-represented single-fold sequences may be masking these signatures, we developed an approach to search both highly diverse protein superfamilies-composed of single-fold and fold-switching variants-and protein subfamilies with more fold-switching variants. This approach successfully revealed coevolution of amino acid pairs uniquely corresponding to both conformations of 56/58 fold-switching proteins from distinct families. Then, using a set of coevolved amino acid pairs predicted by our approach, we successfully biased AlphaFold2 5 to predict two experimentally consistent conformations of a candidate protein with unsolved structure. The discovery of widespread dual-fold coevolution indicates that fold-switching sequences have been preserved by natural selection, implying that their functionalities provide evolutionary advantage and paving the way for predictions of diverse protein structures from single sequences.

DeepHomo2.0: improved protein-protein contact prediction of homodimers by transformer-enhanced deep learning

Protein-protein interactions play an important role in many biological processes. However, although structure prediction for monomer proteins has achieved great progress with the advent of advanced deep learning algorithms like AlphaFold, the structure prediction for protein-protein complexes remains an open question. Taking advantage of the Transformer model of ESM-MSA, we have developed a deep learning-based model, named DeepHomo2.0, to predict protein-protein interactions of homodimeric complexes by leveraging the direct-coupling analysis (DCA) and Transformer features of sequences and the structure features of monomers. DeepHomo2.0 was extensively evaluated on diverse test sets and compared with eight state-of-the-art methods including protein language model-based, DCA-based and machine learning-based methods. It was shown that DeepHomo2.0 achieved a high precision of >70% with experimental monomer structures and >60% with predicted monomer structures for the top 10 predicted contacts on the test sets and outperformed the other eight methods. Moreover, even the version without using structure information, named DeepHomoSeq, still achieved a good precision of >55% for the top 10 predicted contacts. Integrating the predicted contacts into protein docking significantly improved the structure prediction of realistic Critical Assessment of Protein Structure Prediction homodimeric complexes. DeepHomo2.0 and DeepHomoSeq are available at http://huanglab.phys.hust.edu.cn/DeepHomo2/.

Diversifying Design of Nucleic Acid Aptamers Using Unsupervised Machine Learning

Inverse design of short single-stranded RNA and DNA sequences (aptamers) is the task of finding sequences that satisfy a set of desired criteria. Relevant criteria may be, for example, the presence of specific folding motifs, binding to molecular ligands, sensing properties, and so on. Most practical approaches to aptamer design identify a small set of promising candidate sequences using high-throughput experiments (e.g., SELEX) and then optimize performance by introducing only minor modifications to the empirically found candidates. Sequences that possess the desired properties but differ drastically in chemical composition will add diversity to the search space and facilitate the discovery of useful nucleic acid aptamers. Systematic diversification protocols are needed. Here we propose to use an unsupervised machine learning model known as the Potts model to discover new, useful sequences with controllable sequence diversity. We start by training a Potts model using the maximum entropy principle on a small set of empirically identified sequences unified by a common feature. To generate new candidate sequences with a controllable degree of diversity, we take advantage of the model's spectral feature: an "energy" bandgap separating sequences that are similar to the training set from those that are distinct. By controlling the Potts energy range that is sampled, we generate sequences that are distinct from the training set yet still likely to have the encoded features. To demonstrate performance, we apply our approach to design diverse pools of sequences with specified secondary structure motifs in 30-mer RNA and DNA aptamers.

The protein common assembly database (ProtCAD)-a comprehensive structural resource of protein complexes

Proteins often act through oligomeric interactions with other proteins. X-ray crystallography and cryo-electron microscopy provide detailed information on the structures of biological assemblies, defined as the most likely biologically relevant structures derived from experimental data. In crystal structures, the most relevant assembly may be ambiguously determined, since multiple assemblies observed in the crystal lattice may be plausible. It is estimated that 10-15% of PDB entries may have incorrect or ambiguous assembly annotations. Accurate assemblies are required for understanding functional data and training of deep learning methods for predicting assembly structures. As with any other kind of biological data, replication via multiple independent experiments provides important validation for the determination of biological assembly structures. Here we present the Protein Common Assembly Database (ProtCAD), which presents clusters of protein assembly structures observed in independent structure determinations of homologous proteins in the Protein Data Bank (PDB). ProtCAD is searchable by PDB entry, UniProt identifiers, or Pfam domain designations and provides downloads of coordinate files, PyMol scripts, and publicly available assembly annotations for each cluster of assemblies. About 60% of PDB entries contain assemblies in clusters of at least 2 independent experiments. All clusters and coordinates are available on ProtCAD web site (http://dunbrack2.fccc.edu/protcad).