Protein Structure Prediction Research Articles

Accurate and efficient protein embedding using multi-teacher distillation learning.

Protein embedding, which represents proteins as numerical vectors, is a crucial step in various learning-based protein annotation/classification problems, including gene ontology prediction, protein-protein interaction prediction, and protein structure prediction. However, existing protein embedding methods are often computationally expensive due to their large number of parameters, which can reach millions or even billions. The growing availability of large-scale protein datasets and the need for efficient analysis tools have created a pressing demand for efficient protein embedding methods. We propose a novel protein embedding approach based on multi-teacher distillation learning, which leverages the knowledge of multiple pre-trained protein embedding models to learn a compact and informative representation of proteins. Our method achieves comparable performance to state-of-the-art methods while significantly reducing computational costs and resource requirements. Specifically, our approach reduces computational time by ∼70% and maintains ±1.5% accuracy as the original large models. This makes our method well-suited for large-scale protein analysis and enables the bioinformatics community to perform protein embedding tasks more efficiently. The source code of MTDP is available via https://github.com/KennthShang/MTDP.

A genetic investigation in five Chinese families with keratoconus.

This study investigated the genetic characteristics of five Chinese families with keratoconus (KC). In the five families affected by KC, medical records, clinical observations, and blood samples were collected from all individuals. All KC family members (n = 20) underwent both whole exome sequencing of genomic DNA and Sanger sequencing to confirm the variants. Online software was utilized to analyze all variants, and the online server I-TASSER was employed for in silico predictions of the three-dimensional protein structures of the variants. The newly discovered variants and single nucleotide polymorphisms were further examined in 322 sporadic KC patients. The Pentacam tomographic composite index in those affected first-degree family members of the probands showed a pathological change. Five new variants were detected in the five probands and other affected members in their families: a heterozygous missense variant g.19043832C>T (p.Ser145Asn) in the homer scaffolding protein 3 (HOMER3) gene; a heterozygous missense variant g.99452113G>A (p.Gly483Arg) in the insulin-like growth factor 1 receptor (IGF1R) gene; a heterozygous missense variant g.55118280G>T (p.Trp843Leu) in the echinoderm microtubule-associated protein like 6 (EML6) gene; a heterozygous frameshift variant c. 1226_1227del (p.Gln410Glufs*17) in the DOP1 leucine zipper-like protein B (DOP1B) gene; and a heterozygous splice-site variant c.7776+2T>A in the neurobeachin-like protein 2 (NBEAL2) gene. These variations were predicted to be potentially pathogenic and associated with KC. Five novel variants in HOMER3, IGF1R, EML6, DOP1B, and NBEAL2 genes were identified in this study and may be associated with the pathogenesis of KC. This study provides new information about the gene variants and their protein changes in KC patients. The findings should be explored further and could potentially be applied to the early diagnosis of KC before clinical onset.

Development of Drug Discovery Platforms Using Artificial Intelligence and Cheminformatics.

Recently, remarkable progress has been achieved in artificial intelligence (AI), including machine learning. Various AI models have been proposed for drug discovery, including the design of small molecules, activity prediction, and three-dimensional (3D) structure prediction of proteins. AI consists of diverse elements, including information retrieval and machine learning, and can be used in a wide range of drug discovery scenarios. In this review, we focused on AI for small-molecule drug discovery with respect to molecular design, activity prediction, and prediction of the binding poses of compounds to target molecules. We also discussed the applications of AI in academic drug discovery.

Deep learning methods for protein structure prediction

AbstractProtein structure prediction (PSP) has been a prominent topic in bioinformatics and computational biology, aiming to predict protein function and structure from sequence data. The three‐dimensional conformation of proteins is pivotal for their intricate biological roles. With the advancement of computational capabilities and the adoption of deep learning (DL) technologies (especially Transformer network architectures), the PSP field has ushered in a brand‐new era of “neuralization.” Here, we focus on reviewing the evolution of PSP from traditional to modern deep learning‐based approaches and the characteristics of various structural prediction methods. This emphasizes the advantages of deep learning‐based hybrid prediction methods over traditional approaches. This study also provides a summary analysis of widely used bioinformatics databases and the latest structure prediction models. It discusses deep learning networks and algorithmic optimization for model training, validation, and evaluation. In addition, a summary discussion of the major advances in deep learning‐based protein structure prediction is presented. The update of AlphaFold 3 further extends the boundaries of prediction models, especially in protein‐small molecule structure prediction. This marks a key shift toward a holistic approach in biomolecular structure elucidation, aiming at solving almost all sequence‐to‐structure puzzles in various biological phenomena.

Molecular genetic analysis of two novel a allele to cause Ax phenotype in Chinese

BackgroundMutations of ABO gene may cause the dysfunction of ABO glycosyltransferase (GT) that can result in weak ABO phenotypes. Here, we identified two novel weak ABO subgroup alleles and explored the mechanism that caused Ax phenotype. Materials and methodsThe ABO phenotyping and genotyping were performed by serological studies and direct DNA sequencing of ABO gene. The role of the mutations was evaluated by 3D model, predicting protein structure changes, and in vitro expression assay. The total glycosyltransferase transfer capacity in supernatant of transfected cells was examined. ResultsThe results of serological showed the subject RJ23 and RJ52 both were Ax phenotypes. The novel A alleles, Avar-1 and Avar-2 were identified according to the gene analysis. Both Avar-1 and Avar-2 harbored recombinant heterozygous alleles, specifically A2.05 and O.01.02. These alleles showcased substitutions at positions c.106G > T, c.189C > T, c.220C > T, and c.1009A > G in their respective exons. It is worth noting that the crossing-over regions of these two alleles differed from each other. In vitro expression study showed that GTA mutant impaired H to A antigen conversion, and the mutant did not affect the production of GTA though the Western bolt. In silico analysis showed that GTA mutant may change the local conformation and the stability of GT. ConclusionsThe Avar-1 and Avar-2 alleles were identified, which could cause the Ax phenotype through changing the local conformation and reducing stability of the GTA.

Analysis of Phylogenetics and Structural Relationships of the Cytochrome b Gene Protein in Mahseer Fish (Tor sp.)

Abstract The Cytochrome b gene from mitochondrial DNA encodes a protein that can be used as a genetic marker to identify species. This study aims to determine the bioinformatic characteristics of the gene encoding Cyt b isolated from various Mahseer species, Tor tambroides, Tor douronensis, Tor khudree, and Tor putitora. Mahseer Cyt b sequences can be used for detailed computational investigation of the properties of 3D protein models and phylogenetics. The data analysis technique uses nucleotides with a sequence length of 330 bp obtained from NCBI. The MEGA X program was used for phylogenetic analysis. 3D protein structure predictions were obtained through the Swiss model and protein homology program using PYMOL software. The results of phylogenetics show close kinship relationships with the closest genetic distance of 0.000, while the farthest distance is 0.068. The nucleotide composition contains Thymine (30.23%), Cytosine (25.81%), Arginine (28.05%), and Guanine (15.92%). 3D modeling of the Cyt b protein Mahseer referring to the species Neolissochilus hexagonolepis (McClelland, 1839). The 3D protein structure prediction was assessed via a Ramachandran plot of 99.07%. This study can be used as a reference to support sustainable conservation efforts by observing the populations and habitat conditions.

Characterization of 19 novel gene mutation sites associated with autosome-dominant polycystic kidney disease

By analyzing the of genetic testing data of patients with renal polycystic kidney disease and their relatives, this study aims to identify unreported novel gene mutation sites associated with autosomal dominant polycystic kidney disease (ADPKD). Structural prediction software was employed to investigate protein structural changes before and after mutations, explore genotype-phenotype correlations, and enrich the ADPKD gene database. In this single-center retrospective study, patients with multiple renal cysts diagnosed from January 2019 to February 2023 at the Zhong Da Hospital Southeast University were included. Genetic and clinical data of patients and their families were collected. Unreported novel gene mutation sites associated with ADPKD were identified. The AlphaFold v2.3.1 software was used to predict protein structures. Changes in protein structure before and after mutations were compared to explore genotype-phenotype correlations and enrich the ADPKD gene database. Twelve mutated genes associated with renal cysts were detected in 52 families. Nineteen novel gene mutation sites associated with ADPKD were identified, including 17 mutations in the PKD1 gene (one splicing mutation, seven frameshift mutations, four nonsense mutations, one whole-codon insertion, and four missense mutations); one ALG9 missense mutation; and one chromosomal structural variation. Truncating mutations in the PKD1 gene were correlated with a more severe clinical phenotype, while non-truncating mutations were associated with greater clinical heterogeneity. Numerous novel gene mutation sites associated with ADPKD remain unreported. Therefore, it is essential to analyze the pathogenicity of these novel mutation sites, establish genotype-phenotype correlations, and enrich the ADPKD gene database.

The path to the G protein-coupled receptor structural landscape: Major milestones and future directions.

G protein-coupled receptors (GPCRs) play a crucial role in cell function by transducing signals from the extracellular environment to the inside of the cell. They mediate the effects of various stimuli, including hormones, neurotransmitters, ions, photons, food tastants and odorants, and are renowned drug targets. Advancements in structural biology techniques, including X-ray crystallography and cryo-electron microscopy (cryo-EM), have driven the elucidation of an increasing number of GPCR structures. These structures reveal novel features that shed light on receptor activation, dimerization and oligomerization, dichotomy between orthosteric and allosteric modulation, and the intricate interactions underlying signal transduction, providing insights into diverse ligand-binding modes and signalling pathways. However, a substantial portion of the GPCR repertoire and their activation states remain structurally unexplored. Future efforts should prioritize capturing the full structural diversity of GPCRs across multiple dimensions. To do so, the integration of structural biology with biophysical and computational techniques will be essential. We describe in this review the progress of nuclear magnetic resonance (NMR) to examine GPCR plasticity and conformational dynamics, of atomic force microscopy (AFM) to explore the spatial-temporal dynamics and kinetic aspects of GPCRs, and the recent breakthroughs in artificial intelligence for protein structure prediction to characterize the structures of the entire GPCRome. In summary, the journey through GPCR structural biology provided in this review illustrates how far we have come in decoding these essential proteins architecture and function. Looking ahead, integrating cutting-edge biophysics and computational tools offers a path to navigating the GPCR structural landscape, ultimately advancing GPCR-based applications.

An end-to-end framework for the prediction of protein structure and fitness from single sequence

Significant research progress has been made in the field of protein structure and fitness prediction. Particularly, single-sequence-based structure prediction methods like ESMFold and OmegaFold achieve a balance between inference speed and prediction accuracy, showing promise for many downstream prediction tasks. Here, we propose SPIRED, a single-sequence-based structure prediction model that exhibits comparable performance to the state-of-the-art methods but with approximately 5-fold acceleration in inference and at least one order of magnitude reduction in training consumption. By integrating SPIRED with downstream neural networks, we compose an end-to-end framework named SPIRED-Fitness for the rapid prediction of both protein structure and fitness from single sequence with satisfactory accuracy. Moreover, SPIRED-Stab, the derivative of SPIRED-Fitness, achieves state-of-the-art performance in predicting the mutational effects on protein stability.

Integrating Large-Scale Protein Structure Prediction into Human Genetics Research.

The last five years have seen impressive progress in deep learning models applied to protein research. Most notably, sequence-based structure predictions have seen transformative gains in the form of AlphaFold2 and related approaches. Millions of missense protein variants in the human population lack annotations, and these computational methods are a valuable means to prioritize variants for further analysis. Here, we review the recent progress in deep learning models applied to the prediction of protein structure and protein variants, with particular emphasis on their implications for human genetics and health. Improved prediction of protein structures facilitates annotations of the impact of variants on protein stability, protein-protein interaction interfaces, and small-molecule binding pockets. Moreover, it contributes to the study of host-pathogen interactions and the characterization of protein function. As genome sequencing in large cohorts becomes increasingly prevalent, we believe that better integration of state-of-the-art protein informatics technologies into human genetics research is of paramount importance.

Homeodomain leucine zipper (HD-ZIP) gene family in Apostasia shenzhenica Z. J. Liu & L. J. Chen (Orchidaceae): In silico identification and characterization

The homeodomain leucine zipper (HD-ZIP) gene family constitutes an important class of transcription factors found in plants that play crucial roles in various aspects of growth, development and stress regulation. Despite this, there have been minuscule reports on identification and characterization of HD-ZIP gene family in orchids. The present study identifies 31 putative genes (AsHD-ZIP) in Apostasia shenzhenica, using in silico approaches. All of the identified transcription factors were grouped into four subfamilies (HD-ZIP I to IV) based on the distribution of various conserved domains and motifs i.e., HD-ZIP I (homeodomain and leucine zipper), HD-ZIP II (homeodomain, leucine zipper and CPSCE), HD-ZIP III (homeodomain, leucine zipper, MEKHLA and START) and HD-ZIP IV (homeodomain, leucine zipper and START). Prediction of physicochemical properties, protein structures, protein-protein interactions, gene ontology, and gene structures further validated this classification. Phylogenetic analysis showed clustering of all subgroups separately along with their orthologs in Arabidopsis thaliana, Oryza sativa, Dendrobium catenatum and Phalaenopsis equestris. Furthermore, spatio-temporal expression profiling and presence of cis-regulatory elements confirmed tissue specific expression. As HD-ZIP genes are vital players in regulation of plant growth and development as well as in biotic and abiotic stress responses, this study provides a foundation for genetic improvement in orchids through functional validation and characterization of HD-ZIP gene family.

RLpMIEC: High-Affinity Peptide Generation Targeting Major Histocompatibility Complex-I Guided and Interpreted by Interaction Spectrum-Navigated Reinforcement Learning.

Major histocompatibility complex (MHC) plays a vital role in presenting epitopes (short peptides from pathogenic proteins) to T-cell receptors (TCRs) to trigger the subsequent immune responses. Vaccine design targeting MHC generally aims to find epitopes with a high binding affinity for MHC presentation. Nevertheless, to find novel epitopes usually requires high-throughput screening of bulk peptide database, which is time-consuming, labor-intensive, more unaffordable, and very expensive. Excitingly, the past several years have witnessed the great success of artificial intelligence (AI) in various fields, such as natural language processing (NLP, e.g., GPT-4), protein structure prediction and engineering (e.g., AlphaFold2), and so on. Therefore, herein, we propose a deep reinforcement-learning (RL)-based generative algorithm, RLpMIEC, to quantitatively design peptide targeting MHC-I systems. Specifically, RLpMIEC combines the energetic spectrum (namely, the molecular interaction energy component, MIEC) based on the peptide-MHC interaction and the sequence information to generate peptides with strong binding affinity and precise MIEC spectra to accelerate the discovery of candidate peptide vaccines. RLpMIEC performs well in all the generative capability evaluations and can generate peptides with strong binding affinities and precise MIECs and, moreover, with high interpretability, demonstrating its powerful capability in participation for accelerating peptide-based vaccine development.

Evaluation of AlphaFold 3's Protein-Protein Complexes for Predicting Binding Free Energy Changes upon Mutation.

AlphaFold 3 (AF3), the latest version of protein structure prediction software, goes beyond its predecessors by predicting protein-protein complexes. It could revolutionize drug discovery and protein engineering, marking a major step toward comprehensive, automated protein structure prediction. However, independent validation of AF3's predictions is necessary. In this work, we evaluate AF3 complex structures using the SKEMPI 2.0 database which involves 317 protein-protein complexes and 8338 mutations. AF3 complex structures when applied to the most advanced TDL model, MT-TopLap (MultiTask-Topological Laplacian), give rise to a very good Pearson correlation coefficient of 0.86 for predicting protein-protein binding free energy changes upon mutation, which is slightly less than the 0.88 achieved earlier with the Protein Data Bank (PDB) structures. Nonetheless, AF3 complex structures led to a 8.6% increase in the prediction RMSE compared to original PDB complex structures. Additionally, some of AF3's complex structures have large errors, which were not captured in its ipTM performance metric. Finally, it is found that AF3's complex structures are not reliable for intrinsically flexible regions or domains.

Bridging the Gap between Sequence and Structure Classifications of Proteins with AlphaFold Models

Classification of protein domains based on homology and structural similarity serves as a fundamental tool to gain biological insights into protein function. Recent advancements in protein structure prediction, exemplified by AlphaFold, have revolutionized the availability of protein structural data. We focus on classifying about 9000 Pfam families into ECOD (Evolutionary Classification of Domains) by using predicted AlphaFold models and the DPAM (Domain Parser for AlphaFold Models) tool. Our results offer insights into their homologous relationships and domain boundaries. More than half of these Pfam families contain DPAM domains that can be confidently assigned to the ECOD hierarchy. Most assigned domains belong to highly populated folds such as Immunoglobulin-like (IgL), Armadillo (ARM), helix-turn-helix (HTH), and Src homology 3 (SH3). A large fraction of DPAM domains, however, cannot be confidently assigned to ECOD homologous groups. These unassigned domains exhibit statistically different characteristics, including shorter average length, fewer secondary structure elements, and more abundant transmembrane segments. They could potentially define novel families remotely related to domains with known structures or novel superfamilies and folds. Manual scrutiny of a subset of these domains revealed an abundance of internal duplications and recurring structural motifs. Exploring sequence and structural features such as disulfide bond patterns, metal-binding sites, and enzyme active sites helped uncover novel structural folds as well as remote evolutionary relationships. By bridging the gap between sequence-based Pfam and structure-based ECOD domain classifications, our study contributes to a more comprehensive understanding of the protein universe by providing structural and functional insights into previously uncharacterized proteins.

Birth of protein folds and functions in the virome

The rapid evolution of viruses generates proteins that are essential for infectivity and replication but with unknown functions, due to extreme sequence divergence1. Here, using a database of 67,715 newly predicted protein structures from 4,463 eukaryotic viral species, we found that 62% of viral proteins are structurally distinct and lack homologues in the AlphaFold database2,3. Among the remaining 38% of viral proteins, many have non-viral structural analogues that revealed surprising similarities between human pathogens and their eukaryotic hosts. Structural comparisons suggested putative functions for up to 25% of unannotated viral proteins, including those with roles in the evasion of innate immunity. In particular, RNA ligase T-like phosphodiesterases were found to resemble phage-encoded proteins that hydrolyse the host immune-activating cyclic dinucleotides 3′,3′- and 2′,3′-cyclic GMP-AMP (cGAMP). Experimental analysis showed that RNA ligase T homologues encoded by avian poxviruses similarly hydrolyse cGAMP, showing that RNA ligase T-mediated targeting of cGAMP is an evolutionarily conserved mechanism of immune evasion that is present in both bacteriophage and eukaryotic viruses. Together, the viral protein structural database and analyses presented here afford new opportunities to identify mechanisms of virus–host interactions that are common across the virome.

AlphaFold and Protein Folding: Not Dead Yet! The Frontier Is Conformational Ensembles.

Like the black knight in the classic Monty Python movie, grand scientific challenges such as protein folding are hard to finish off. Notably, AlphaFold is revolutionizing structural biology by bringing highly accurate structure prediction to the masses and opening up innumerable new avenues of research. Despite this enormous success, calling structure prediction, much less protein folding and related problems, "solved" is dangerous, as doing so could stymie further progress. Imagine what the world would be like if we had declared flight solved after the first commercial airlines opened and stopped investing in further research and development. Likewise, there are still important limitations to structure prediction that we would benefit from addressing. Moreover, we are limited in our understanding of the enormous diversity of different structures a single protein can adopt (called a conformational ensemble) and the dynamics by which a protein explores this space. What is clear is that conformational ensembles are critical to protein function, and understanding this aspect of protein dynamics will advance our ability to design new proteins and drugs.

Predicting protein conformational motions using energetic frustration analysis and AlphaFold2

Proteins perform their biological functions through motion. Although high throughput prediction of the three-dimensional static structures of proteins has proved feasible using deep-learning-based methods, predicting the conformational motions remains a challenge. Purely data-driven machine learning methods encounter difficulty for addressing such motions because available laboratory data on conformational motions are still limited. In this work, we develop a method for generating protein allosteric motions by integrating physical energy landscape information into deep-learning-based methods. We show that local energetic frustration, which represents a quantification of the local features of the energy landscape governing protein allosteric dynamics, can be utilized to empower AlphaFold2 (AF2) to predict protein conformational motions. Starting from ground state static structures, this integrative method generates alternative structures as well as pathways of protein conformational motions, using a progressive enhancement of the energetic frustration features in the input multiple sequence alignment sequences. For a model protein adenylate kinase, we show that the generated conformational motions are consistent with available experimental and molecular dynamics simulation data. Applying the method to another two proteins KaiB and ribose-binding protein, which involve large-amplitude conformational changes, can also successfully generate the alternative conformations. We also show how to extract overall features of the AF2 energy landscape topography, which has been considered by many to be black box. Incorporating physical knowledge into deep-learning-based structure prediction algorithms provides a useful strategy to address the challenges of dynamic structure prediction of allosteric proteins.

Advances in the Application of Protein Language Modeling for Nucleic Acid Protein Binding Site Prediction.

Protein and nucleic acid binding site prediction is a critical computational task that benefits a wide range of biological processes. Previous studies have shown that feature selection holds particular significance for this prediction task, making the generation of more discriminative features a key area of interest for many researchers. Recent progress has shown the power of protein language models in handling protein sequences, in leveraging the strengths of attention networks, and in successful applications to tasks such as protein structure prediction. This naturally raises the question of the applicability of protein language models in predicting protein and nucleic acid binding sites. Various approaches have explored this potential. This paper first describes the development of protein language models. Then, a systematic review of the latest methods for predicting protein and nucleic acid binding sites is conducted by covering benchmark sets, feature generation methods, performance comparisons, and feature ablation studies. These comparisons demonstrate the importance of protein language models for the prediction task. Finally, the paper discusses the challenges of protein and nucleic acid binding site prediction and proposes possible research directions and future trends. The purpose of this survey is to furnish researchers with actionable suggestions for comprehending the methodologies used in predicting protein-nucleic acid binding sites, fostering the creation of protein-centric language models, and tackling real-world obstacles encountered in this field.

Origins of life: The Protein Folding Problem all over again?

How did specific useful protein sequences arise from simpler molecules at the origin of life? This seemingly needle-in-a-haystack problem has remarkably close resemblance to the old Protein Folding Problem, for which the solution is now known from statistical physics. Based on the logic that Origins must have come only after there was an operative evolution mechanism—which selects on phenotype, not genotype—we give a perspective that proteins and their folding processes are likely to have been the primary driver of the early stages of the origin of life.

Mutations in the WG and GW motifs of the three RNA silencing suppressors of grapevine fanleaf virus alter their systemic suppression ability and affect virus infectivity.

Viral suppressors of RNA silencing (VSRs) encoded by grapevine fanleaf virus (GFLV), one of the most economically consequential viruses of grapevine (Vitis spp.), were recently identified. GFLV VSRs include the RNA1-encoded protein 1A and the putative helicase protein 1BHel, as well as their fused form (1ABHel). Key characteristics underlying the suppression function of the GFLV VSRs are unknown. In this study, we explored the role of the conserved tryptophan-glycine (WG) motif in protein 1A and glycine-tryptophan (GW) motif in protein 1BHel in their systemic RNA silencing suppression ability by co-infiltrating Nicotiana benthamiana 16c line plants with a GFP silencing construct and a wildtype or a mutant GFLV VSR. We analyzed and compared wildtype and mutant GFLV VSRs for their (i) efficiency at suppressing RNA silencing, (ii) ability to limit siRNA accumulation, (iii) modulation of the expression of six host genes involved in RNA silencing, (iv) impact on virus infectivity in planta, and (v) variations in predicted protein structures using molecular and biochemical assays, as well as bioinformatics tools such as AlphaFold2. Mutating W to alanine (A) in WG of proteins 1A and 1ABHel abolished their ability to induce systemic RNA silencing suppression, limit siRNA accumulation, and downregulate NbAGO2 expression by 1ABHel. This mutation in the GFLV genome resulted in a non-infectious virus. Mutating W to A in GW of proteins 1BHel and 1ABHel reduced their ability to suppress systemic RNA silencing and abolished the downregulation of NbDCL2, NbDCL4,, and NbRDR6 expression by 1BHel. This mutation in the GFLV genome delayed infection at the local level and inhibited systemic infection in planta. Double mutations of W to A in WG and GW of protein 1ABHel abolished its ability to induce RNA silencing suppression, limit siRNA accumulation, and downregulate NbDCL2 and NbRDR6 expression. Finally, in silico protein structure prediction indicated that a W to A substitution potentially modifies the structure and physicochemical properties of the three GFLV VSRs. Together, this study provided insights into the specific roles of WG/GW not only in GFLV VSR functions but also in GFLV biology.