Accurate Protein Structure Prediction Research Articles

Easy and accurate protein structure prediction using ColabFold.

Since its public release in 2021, AlphaFold2 (AF2) has made investigating biological questions, by using predicted protein structures of single monomers or full complexes, a common practice. ColabFold-AF2 is an open-source Jupyter Notebook inside Google Colaboratory and a command-line tool that makes it easy to use AF2 while exposing its advanced options. ColabFold-AF2 shortens turnaround times of experiments because of its optimized usage of AF2's models. In this protocol, we guide the reader through ColabFold best practices by using three scenarios: (i) monomer prediction, (ii) complex prediction and (iii) conformation sampling. The first two scenarios cover classic static structure prediction and are demonstrated on the human glycosylphosphatidylinositol transamidase protein. The third scenario demonstrates an alternative use case of the AF2 models by predicting two conformations of the human alanine serine transporter 2. Users can run the protocol without computational expertise via Google Colaboratory or in a command-line environment for advanced users. Using Google Colaboratory, it takes <2 h to run each procedure. The data and code for this protocol are available at https://protocol.colabfold.com .

The Predicted Structure of a Thermophilic Malate Synthase

This project aims to solve the structure of the crenarchaeal Sulfolobus acidocaldarius enzyme malate synthase. Other known malate synthase enzymes have been found to require a magnesium ion in the active site to carry out catalytic activities, but a study reported that S. acidocaldarius malate synthase does not require magnesium. This suggests a novel mechanism for this enzyme. Additionally, the mature S. acidocaldarius protein is approximately 100 residues larger than any other structurally characterized malate synthase. It has also been reported to form a dimer, while previously solved structures have only displayed monomeric, trimeric, and hexameric arrangements. We plan to determine the structure experimentally. However, major advances in the accuracy of protein structure prediction were made recently by AlphaFold, an artificial intelligence system developed by DeepMind, which has revolutionized the field and has largely solved the protein folding problem. A similar AI system, RoseTTAFold, developed by David Baker’s lab at the University of Washington, has been publicly available. Here, we report our analysis of the structure of this protein, predicted using both of these algorithms and of a predicted structural model for the dimeric form of the enzyme using ClusPro. Our results strongly support a conserved catalytic mechanism requiring magnesium, which is common with all previously solved malate synthase isoforms. KEYWORDS: Glyoxylate Cycle; Malate synthase; Protein Prediction; Thermophile; Sulfolobus acidocaldarius; Magnesium; AlphaFold; RoseTTAFold

High-throughput prediction of protein conformational distributions with subsampled AlphaFold2

This paper presents an innovative approach for predicting the relative populations of protein conformations using AlphaFold 2, an AI-powered method that has revolutionized biology by enabling the accurate prediction of protein structures. While AlphaFold 2 has shown exceptional accuracy and speed, it is designed to predict proteins’ ground state conformations and is limited in its ability to predict conformational landscapes. Here, we demonstrate how AlphaFold 2 can directly predict the relative populations of different protein conformations by subsampling multiple sequence alignments. We tested our method against nuclear magnetic resonance experiments on two proteins with drastically different amounts of available sequence data, Abl1 kinase and the granulocyte-macrophage colony-stimulating factor, and predicted changes in their relative state populations with more than 80% accuracy. Our subsampling approach worked best when used to qualitatively predict the effects of mutations or evolution on the conformational landscape and well-populated states of proteins. It thus offers a fast and cost-effective way to predict the relative populations of protein conformations at even single-point mutation resolution, making it a useful tool for pharmacology, analysis of experimental results, and predicting evolution.

Solving protein structures by combining structure prediction, molecular replacement and direct-methods-aided model completion.

Highly accurate protein structure prediction can generate accurate models of protein and protein-protein complexes in X-ray crystallography. However, the question of how to make more effective use of predicted models for completing structure analysis, and which strategies should be employed for the more challenging cases such as multi-helical structures, multimeric structures and extremely large structures, both in the model preparation and in the completion steps, remains open for discussion. In this paper, a new strategy is proposed based on the framework of direct methods and dual-space iteration, which can greatly simplify the pre-processing steps of predicted models both in normal and in challenging cases. Following this strategy, full-length models or the conservative structural domains could be used directly as the starting model, and the phase error and the model bias between the starting model and the real structure would be modified in the direct-methods-based dual-space iteration. Many challenging cases (from CASP14) have been tested for the general applicability of this constructive strategy, and almost complete models have been generated with reasonable statistics. The hybrid strategy therefore provides a meaningful scheme for X-ray structure determination using a predicted model as the starting point.

Unsupervisedly Prompting AlphaFold2 for Accurate Few-Shot Protein Structure Prediction.

Data-driven predictive methods that can efficiently and accurately transform protein sequences into biologically active structures are highly valuable for scientific research and medical development. Determining an accurate folding landscape using coevolutionary information is fundamental to the success of modern protein structure prediction methods. As the state of the art, AlphaFold2 has dramatically raised the accuracy without performing explicit coevolutionary analysis. Nevertheless, its performance still shows strong dependence on available sequence homologues. Based on the interrogation on the cause of such dependence, we presented EvoGen, a meta generative model, to remedy the underperformance of AlphaFold2 for poor MSA targets. By prompting the model with calibrated or virtually generated homologue sequences, EvoGen helps AlphaFold2 fold accurately in the low-data regime and even achieve encouraging performance with single-sequence predictions. Being able to make accurate predictions with few-shot MSA not only generalizes AlphaFold2 better for orphan sequences but also democratizes its use for high-throughput applications. Besides, EvoGen combined with AlphaFold2 yields a probabilistic structure generation method that could explore alternative conformations of protein sequences, and the task-aware differentiable algorithm for sequence generation will benefit other related tasks including protein design.

Accurate prediction of protein folding mechanisms by simple structure-based statistical mechanical models

Recent breakthroughs in highly accurate protein structure prediction using deep neural networks have made considerable progress in solving the structure prediction component of the ‘protein folding problem’. However, predicting detailed mechanisms of how proteins fold into specific native structures remains challenging, especially for multidomain proteins constituting most of the proteomes. Here, we develop a simple structure-based statistical mechanical model that introduces nonlocal interactions driving the folding of multidomain proteins. Our model successfully predicts protein folding processes consistent with experiments, without the limitations of protein size and shape. Furthermore, slight modifications of the model allow prediction of disulfide-oxidative and disulfide-intact protein folding. These predictions depict details of the folding processes beyond reproducing experimental results and provide a rationale for the folding mechanisms. Thus, our physics-based models enable accurate prediction of protein folding mechanisms with low computational complexity, paving the way for solving the folding process component of the ‘protein folding problem’.

Struct2GO: protein function prediction based on Graph pooling algorithm and AlphaFold2 structure information.

In recent years, there has been a breakthrough in protein structure prediction, and the AlphaFold2 model of the DeepMind team has improved the accuracy of protein structure prediction to the atomic level. Currently, deep learning-based protein function prediction models usually extract features from protein sequences and combine them with protein-protein interaction networks to achieve good results. However, for newly sequenced proteins that are not in the protein-protein interaction network, such models cannot make effective predictions. To address this, this paper proposes the Struct2GO model, which combines protein structure and sequence data to enhance the precision of protein function prediction and the generality of the model. We obtain amino acid residue embeddings in protein structure through graph representation learning, utilize the graph pooling algorithm based on a self-attention mechanism to obtain the whole graph structure features, and fuse them with sequence features obtained from the protein language model. The results demonstrate that compared with the traditional protein sequence-based function prediction model, the Struct2GO model achieves better results. https://github.com/lyjps/Struct2GO. Supplementary data are available at Bioinformatics online.

In Silico Structural Analysis Exploring Conformational Folding of Protein Variants in Alzheimer's Disease.

Accurate protein structure prediction using computational methods remains a challenge in molecular biology. Recent advances in AI-powered algorithms provide a transformative effect in solving this problem. Even though AlphaFold's performance has improved since its release, there are still limitations that apply to its efficacy. In this study, a selection of proteins related to the pathology of Alzheimer's disease was modeled, with Presenilin-1 (PSN1) and its mutated variants in the foreground. Their structural predictions were evaluated using the ColabFold implementation of AlphaFold, which utilizes MMseqs2 for the creation of multiple sequence alignments (MSAs). A higher number of recycles than the one used in the AlphaFold DB was selected, and no templates were used. In addition, prediction by RoseTTAFold was also applied to address how structures from the two deep learning frameworks match reality. The resulting conformations were compared with the corresponding experimental structures, providing potential insights into the predictive ability of this approach in this particular group of proteins. Furthermore, a comprehensive examination was performed on features such as predicted regions of disorder and the potential effect of mutations on PSN1. Our findings consist of highly accurate superpositions with little or no deviation from experimentally determined domain-level models.

Hidden Glutathione Transferases in the Human Genome.

With the development of accurate protein structure prediction algorithms, artificial intelligence (AI) has emerged as a powerful tool in the field of structural biology. AI-based algorithms have been used to analyze large amounts of protein sequence data including the human proteome, complementing experimental structure data found in resources such as the Protein Data Bank. The EBI AlphaFold Protein Structure Database (for example) contains over 230 million structures. In this study, these data have been analyzed to find all human proteins containing (or predicted to contain) the cytosolic glutathione transferase (cGST) fold. A total of 39 proteins were found, including the alpha-, mu-, pi-, sigma-, zeta- and omega-class GSTs, intracellular chloride channels, metaxins, multisynthetase complex components, elongation factor 1 complex components and others. Three broad themes emerge: cGST domains as enzymes, as chloride ion channels and as protein-protein interaction mediators. As the majority of cGSTs are dimers, the AI-based structure prediction algorithm AlphaFold-multimer was used to predict structures of all pairwise combinations of these cGST domains. Potential homo- and heterodimers are described. Experimental biochemical and structure data is used to highlight the strengths and limitations of AI-predicted structures.

Leveraging Fine-Tuned Language Models in Bioinformatics: A Research Perspective

Bioinformatics, an interdisciplinary field combining biology, computer science, and statistics, has advanced with deep learning and natural language processing techniques. This perspective explores the applications of fine-tuned language models in bioinformatics, highlighting their potential in various domains while discussing challenges and limitations. Fine-tuned language models benefit biomedical literature analysis, extracting information from scientific papers to synthesize knowledge and generate synthetic sequences for DNA, RNA, and protein research. In drug discovery, these models can identify novel drug targets, accelerate virtual screening, and aid drug repurposing by finding new therapeutic indications for existing drugs. For clinical decision support, fine-tuned language models can analyse patient data, medical literature, and guidelines to provide personalized recommendations and alerts to healthcare professionals. They can also aid accurate protein structure prediction for drug design and target identification. In pharmacovigilance, these models can analyse unstructured data sources to detect adverse events from social media, patient forums, and health records, enabling early intervention and improving patient safety. However, challenges like data availability, domain-specific knowledge, bias, interpretability, resource efficiency, ethics, and validation must be addressed for reliable application. Addressing these challenges will unlock the full potential of fine-tuned language models in bioinformatics, driving advancements and benefiting human health. Collaboration between computational and experimental biologists, ethicists, and regulatory bodies is crucial to establish ethical guidelines and best practices for their use.

Predicting Protein-Peptide Interactions: Benchmarking Deep Learning Techniques and a Comparison with Focused Docking.

The accurate prediction of protein structures achieved by deep learning (DL) methods is a significant milestone and has deeply impacted structural biology. Shortly after its release, AlphaFold2 has been evaluated for predicting protein-peptide interactions and shown to significantly outperform RoseTTAfold as well as a conventional blind docking method: PIPER-FlexPepDock. Since then, new AlphaFold2 models, trained specifically to predict multimeric assemblies, have been released and a new ab initio folding model OmegaFold has become available. Here, we assess docking success rates for these new DL folding models and compare their performance with our state-of-the-art, focused peptide-docking software AutoDock CrankPep (ADCP). The evaluation is done using the same dataset and performance metric for all methods. We show that, for a set of 99 nonredundant protein-peptide complexes, the new AlphaFold2 model outperforms other Deep Learning approaches and achieves remarkable docking success rates for peptides. While the docking success rate of ADCP is more modest when considering the top-ranking solution only, it samples correct solutions for around 62% of the complexes. Interestingly, different methods succeed on different complexes, and we describe a consensus docking approach using ADCP and AlphaFold2, which achieves a remarkable 60% for the top-ranking results and 66% for the top 5 results for this set of 99 protein-peptide complexes.

Combining Protein Conformational Diversity and Phylogenetic Information Using CoDNaS and CoDNaS-Q.

CoDNaS (http://ufq.unq.edu.ar/codnas/) and CoDNaS-Q (http://ufq.unq.edu.ar/codnasq) are repositories of proteins with different degrees of conformational diversity. Following the ensemble nature of the native state, conformational diversity represents the structural differences between the conformers in the ensemble. Each entry in CoDNaS and CoDNaS-Q contains a redundant collection of experimentally determined conformers obtained under different conditions. These conformers represent snapshots of the protein dynamism. While CoDNaS contains examples of conformational diversity at the tertiary level, a recent development, CoDNaS-Q, contains examples at the quaternary level. In the emerging age of accurate protein structure prediction by machine learning approaches, many questions remain open regarding the characterization of protein dynamism. In this context, most bioinformatics resources take advantage of distinct features derived from protein alignments, however, the complexity and heterogeneity of information makes it difficult to recover reliable biological signatures. Here we present five protocols to explore tertiary and quaternary conformational diversity at the individual protein level as well as for the characterization of the distribution of conformational diversity at the protein family level in a phylogenetic context. These protocols can provide curated protein families with experimentally known conformational diversity, facilitating the exploration of sequence determinants of protein dynamism. © 2023 Wiley Periodicals LLC. Basic Protocol 1: Assessing conformational diversity with CoDNaS Alternate Protocol 1: Assessing conformational diversity at the quaternary level with CoDNaS-Q Basic Protocol 2: Exploring conformational diversity in a protein family Alternate Protocol 2: Exploring quaternary conformational diversity in a protein family Basic Protocol 3: Representing conformational diversity in a phylogenetic context.

Prediction of protein structures, functions and interactions using the IntFOLD7, MultiFOLD and ModFOLDdock servers.

The IntFOLD server based at the University of Reading has been a leading method over the past decade in providing free access to accurate prediction of protein structures and functions. In a post-AlphaFold2 world, accurate models of tertiary structures are widely available for even more protein targets, so there has been a refocus in the prediction community towards the accurate modelling of protein-ligand interactions as well as modelling quaternary structure assemblies. In this paper, we describe the latest improvements to IntFOLD, which maintains its competitive structure prediction performance by including the latest deep learning methods while also integrating accurate model quality estimates and 3D models of protein-ligand interactions. Furthermore, we also introduce our two new server methods: MultiFOLD for accurately modelling both tertiary and quaternary structures, with performancewhich has been independently verified to outperform the standard AlphaFold2 methods, and ModFOLDdock, which provides world-leading quality estimates for quaternary structure models. The IntFOLD7, MultiFOLD and ModFOLDdock servers are available at: https://www.reading.ac.uk/bioinf/.

Protein structure prediction using the evolutionary algorithm USPEX.

Protein structure prediction is one of major problems of modern biophysics: current attempts to predict the tertiary protein structure from amino acid sequence are successful mostly when the use of big data and machine learning allows one to reduce the "prediction problem" to the "problem of recognition". Compared with recent successes of deep learning, classical predictive methods lag behind in their accuracy for the prediction of stable conformations. Therefore, in this work we extended the evolutionary algorithm USPEX to predict protein structure based on global optimization starting with the amino acid sequence. Moreover, we compared frequently used force fields for the task of protein structure prediction. Protein structure relaxation and energy calculations were performed using Tinker (with several different force fields) and Rosetta (with REF2015 force field) codes. To create new protein structure models in the USPEX algorithm, we developed novel variation operators. The test of the new method on seven proteins having (for simplicity) no cis-proline (with ω ≈ 0°) residues, and a length of up to 100 residues, revealed that our algorithm predicts tertiary structures of proteins with high accuracy. The comparison of the final potential energies of the predicted protein structures obtained using the USPEX and the Rosetta Abinitio approach showed that in most cases the developed algorithm found structures with close or even lower energy (Amber/Charmm/Oplsaal) and scoring function (REF2015). While USPEX has clearly demonstrated its ability to find very deep energy minima, our study showed that the existing force fields are not sufficiently accurate for accurate blind prediction of protein structures without further experimental verification.

Recent Advances in Biomolecular Structure Modeling and Validation using Deep Learning

Since its appearance in 2021, Alphafold2, an accurate protein structure prediction method, has quickly been adopted and substantially impacted biology research, particularly in the field of structural biology. In this review, we discuss recent advancements in the development of computational structural modeling tools related to Alphafold2. Additionally, we discuss deep learning-based structure modeling methods for cryo-electron microscopy.

Mapping the FtsQBL divisome components in bacterial NTD pathogens as potential drug targets.

Cytokinesis is an essential process in bacterial cell division, and it involves more than 25 essential/non-essential cell division proteins that form a protein complex known as a divisome. Central to the divisome are the proteins FtsB and FtsL binding to FtsQ to form a complex FtsQBL, which helps link the early proteins with late proteins. The FtsQBL complex is highly conserved as a component across bacteria. Pathogens like Vibrio cholerae, Mycobacterium ulcerans, Mycobacterium leprae, and Chlamydia trachomatis are the causative agents of the bacterial Neglected Tropical Diseases Cholera, Buruli ulcer, Leprosy, and Trachoma, respectively, some of which seemingly lack known homologs for some of the FtsQBL complex proteins. In the absence of experimental characterization, either due to insufficient resources or the massive increase in novel sequences generated from genomics, functional annotation is traditionally inferred by sequence similarity to a known homolog. With the advent of accurate protein structure prediction methods, features both at the fold level and at the protein interaction level can be used to identify orthologs that cannot be unambiguously identified using sequence similarity methods. Using the FtsQBL complex proteins as a case study, we report potential remote homologs using Profile Hidden Markov models and structures predicted using AlphaFold. Predicted ortholog structures show conformational similarity with corresponding E. coli proteins irrespective of their level of sequence similarity. Alphafold multimer was used to characterize remote homologs as FtsB or FtsL, when they were not sufficiently distinguishable at both the sequence or structure level, as their interactions with FtsQ and FtsW play a crucial role in their function. The structures were then analyzed to identify functionally critical regions of the proteins consistent with their homologs and delineate regions potentially useful for inhibitor discovery.

Contact-Assisted Threading in Low-Homology Protein Modeling.

The ability to successfully predict the three-dimensional structure of a protein from its amino acid sequence has made considerable progress in the recent past. The progress is propelled by the improved accuracy of deep learning-based inter-residue contact map predictors coupled with the rising growth of protein sequence databases. Contact map encodes interatomic interaction information that can be exploited for highly accurate prediction of protein structures via contact map threading even for the query proteins that are not amenable to direct homology modeling. As such, contact-assisted threading has garnered considerable research effort. In this chapter, we provide an overview of existing contact-assisted threading methods while highlighting the recent advances and discussing some of the current limitations and future prospects in the application of contact-assisted threading for improving the accuracy of low-homology protein modeling.

Fast and accurate Ab Initio Protein structure prediction using deep learning potentials.

Despite the immense progress recently witnessed in protein structure prediction, the modeling accuracy for proteins that lack sequence and/or structure homologs remains to be improved. We developed an open-source program, DeepFold, which integrates spatial restraints predicted by multi-task deep residual neural-networks along with a knowledge-based energy function to guide its gradient-descent folding simulations. The results on large-scale benchmark tests showed that DeepFold creates full-length models with accuracy significantly beyond classical folding approaches and other leading deep learning methods. Of particular interest is the modeling performance on the most difficult targets with very few homologous sequences, where DeepFold achieved an average TM-score that was 40.3% higher than trRosetta and 44.9% higher than DMPfold. Furthermore, the folding simulations for DeepFold were 262 times faster than traditional fragment assembly simulations. These results demonstrate the power of accurately predicted deep learning potentials to improve both the accuracy and speed of ab initio protein structure prediction.

Are Deep Learning Structural Models Sufficiently Accurate for Free-Energy Calculations? Application of FEP+ to AlphaFold2-Predicted Structures.

The availability of AlphaFold2 has led to great excitement in the scientific community─particularly among drug hunters─due to the ability of the algorithm to predict protein structures with high accuracy. However, beyond globally accurate protein structure prediction, it remains to be determined whether ligand binding sites are predicted with sufficient accuracy in these structures to be useful in supporting computationally driven drug discovery programs. We explored this question by performing free-energy perturbation (FEP) calculations on a set of well-studied protein-ligand complexes, where AlphaFold2 predictions were performed by removing all templates with >30% identity to the target protein from the training set. We observed that in most cases, the ΔΔG values for ligand transformations calculated with FEP, using these prospective AlphaFold2 structures, were comparable in accuracy to the corresponding calculations previously carried out using crystal structures. We conclude that under the right circumstances, AlphaFold2-modeled structures are accurate enough to be used by physics-based methods such as FEP in typical lead optimization stages of a drug discovery program.

Accurate prediction of protein structures and interactions using a three-track neural network

Accurate prediction of protein structures and interactions using a three-track neural network