3D Protein Structure Prediction Research Articles

Molecular Dynamics of Outer Membrane-Embedded Polysaccharide Secretion Porins Reveals Closed Resting-State Surface Gates Targetable by Virtual Fragment Screening for Drug Hotspot Identification.

Recent advances in iterative neural network analyses (e.g., AlphaFold2 and RoseTTA fold) have been revolutionary for protein 3D structure prediction, especially for difficult-to-manipulate α-helical/β-barrel integral membrane proteins. These model structures are calculated based on the coevolution of amino acids within the protein of interest and similarities to existing protein structures; the local effects of the membrane on folding and stability of the calculated model structures are not considered. We recently reported the discovery, 3D modeling, and characterization of 18-β-stranded outer-membrane (OM) WzpX, WzpS, and WzpB β-barrel secretion porins for the exopolysaccharide (EPS), major spore coat polysaccharide (MASC), and biosurfactant polysaccharide (BPS) pathways (respectively) in the Gram-negative social predatory bacterium Myxococcus xanthus DZ2. However, information was not obtained regarding the dynamic behavior of surface-gating WzpX/S/B loop domains or on potential treatments to inactivate these porins. Herein, we developed a molecular dynamics (MD) protocol to study the core stability and loop dynamism of neural network-based integral membrane protein structure models embedded in an asymmetric OM bilayer, using the M. xanthus WzpX, WzpS, and WzpB proteins as test candidates. This was accomplished through integration of the CHARMM-graphical user interface (GUI) and Molecular Operating Environment (MOE) workflows to allow for a rapid simulation system setup and facilitate data analysis. In addition to serving as a method of model structure validation, our molecular dynamics simulations revealed a minimal movement of extracellular WzpX/S/B loops in the absence of an external stimulus as well as druggable cavities between the loops. Virtual screening of a commercial fragment library against these cavities revealed putative fragment-binding hotspots on the cell-surface face of each β-barrel, along with key interacting residues, and identified promising hits for the design of potential binders capable of plugging the β-barrels and inhibiting polysaccharide secretion.

Delving into the significance of the His289Tyr single-nucleotide polymorphism in the glutamate ionotropic receptor kainate-1 (Grik1) gene of a genetically audiogenic seizure model.

Genetic abnormalities affecting glutamate receptors are central to excitatory overload-driven neuronal mechanisms that culminate in seizures, making them pivotal targets in epilepsy research. Increasingly used to advance this field, the genetically audiogenic seizure hamster from Salamanca (GASH/Sal) exhibits generalized seizures triggered by high-intensity acoustic stimulation and harbors significant genetic variants recently identified through whole-exome sequencing. Here, we addressed the influence of the missense single-nucleotide polymorphism (C9586732T, p.His289Tyr) in the glutamate receptor ionotropic kainate-1 (Grik1) gene and its implications for the GASH/Sal seizure susceptibility. Using a protein 3D structure prediction, we showed a potential effect of this sequence variation, located in the amino-terminal domain, on the stability and/or conformation of the kainate receptor subunit-1 protein (GluK1). We further employed a multi-technique approach, encompassing gene expression analysis (RT-qPCR), Western blotting, and immunohistochemistry in bright-field and confocal fluorescence microscopy, to investigate critical seizure-associated brain regions in GASH/Sal animals under seizure-free conditions compared to matched wild-type controls. We detected disruptions in the transcriptional profile of the Grik1 gene within the audiogenic seizure-associated neuronal network. Alterations in GluK1 protein levels were also observed in various brain structures, accompanied by an unexpected lower molecular weight band in the inferior and superior colliculi. This correlated with substantial disparities in GluK1-immunolabeling distribution across multiple brain regions, including the cerebellum, hippocampus, subdivisions of the inferior and superior colliculi, and the prefrontal cortex. Notably, the diffuse immunolabeling accumulated within perikarya, axonal fibers and terminals, exhibiting a prominent concentration in proximity to the cell nucleus. This suggests potential disturbances in the GluK1-trafficking mechanism, which could subsequently affect glutamate synaptic transmission. Overall, our study sheds light on the genetic underpinnings of seizures and underscores the importance of investigating the molecular mechanisms behind synaptic dysfunction in epileptic neural networks, laying a crucial foundation for future research and therapeutic strategies targeting GluK1-containing kainate receptors.

Sequence identification and expression characterization of leptin in the spotted scat, Scatophagus argus.

Scatophagus argus is an important marine culture fish in South and South-East Asia, including Southeast coastal areas of China. Artificial propagation technology for S. argus is not optimum; thus further studies on its reproduction biology are required. Although previous studies have shown that leptin (Lep) can regulate fish reproduction, the role of lep genes in S. argus is unknown. Herein, in silico analysis showed that S. argus has two lep genes (lepa and lepb). Protein 3D-structure prediction showed that Lepa has four α-helices (similar to mammals), while Lepb only has three. Tissue distribution analysis showed that lepa is highly expressed in the liver, whereas lepb was not detected in any tissue. Notably, lepr was expressed in all tissues. Lepa mRNA expression levels in the liver and serum Lep, estradiol (E2) and vitellogenin (Vtg) levels of female fish were significantly higher in ovaries at stage IV than in ovaries at stage II. Serum E2 levels were significantly positively correlated with Vtg levels in female fish at different development stages, while serum E2 was not correlated with Lep levels. Consistently, in vitro incubation of the liver with E2 significantly up-regulated vtga, while it did not affect lepa expression. Recombinant Lep (10nM) significantly up-regulated chicken gonadotropin-releasing hormone (cGnRH/GnRH-II) in the hypothalamus and GnRH receptor (GnRHR) and luteinizing hormone beta (Lhb) in the pituitary. These results suggest that lepa regulates female reproduction in S. argus.

Identification of passion fruit HSF gene family and the functional analysis of PeHSF-C1a in response to heat and osmotic stress

Heat stress transcription factors (HSFs) are the major regulators of plant response to environmental stress, especially heat and drought stress. To gain a deeper understanding of the mechanisms underlying HSFs in the abiotic stress response of passion fruit, we conducted an in silico analysis of the HSF gene family. Through bioinformatics and phylogenetic analyses, we identified 18 PeHSF members and classified them into A, B, and C groups. Collinearity analysis results revealed that the expansion of the PeHSF gene family was due to the presence of segmental duplication. Furthermore, gene structure and protein domain analysis illustrated that PeHSFs in the same subgroup are relatively conserved. Conserved motif and function domain analysis suggested that PeHSF proteins possess typical conserved functional domains of the HSF family. A protein interaction network and 3D structure prediction were used to study the potential regulatory relationship of PeHSFs. Additionally, the subcellular localization results of PeHSF-A6a, PeHSF-B4b, and PeHSF-C1a were consistent with the predictions. RNA-seq and RT-qPCR analysis revealed the expression patterns of PeHSFs in different tissues of passion fruit floral organs. Promoter analysis and the expression patterns of the PeHSFs under different treatments demonstrated their involvement in various abiotic stress processes. Notably, overexpression of PeHSF-C1a consistently enhanced tolerance to drought and heat stress in Arabidopsis. Overall, our findings provide a scientific basis for further functional studies of PeHSFs that could contribute to improvement of passion fruit breeding.

A novel missense mutation of the STK11 gene in a Chinese family with Peutz-Jeghers syndrome

BackgroundPeutz-Jeghers syndrome (PJS) is a rare autosomal dominant inherited disease caused by mutations in the Serine-Threonine Kinase 11 (STK11) gene. This study aimed to diagnose a Chinese pedigree with PJS and to expand the spectrum of STK11 variants.MethodsWe performed an inductive analysis of clinical features, gastrointestinal endoscopy, radiologic imaging, and pathological findings in a Chinese family with PJS. Whole-exome sequencing (WES), Sanger sequencing, and STK11 protein 3D structure prediction were performed for establishing a molecular diagnosis.ResultsThe proband, her mother, and grandfather presented with pigmentation spots on lips, oral mucosa, and fingers. Her mother and grandfather also had pigmentation spots on face and feet, while her brother had pigmentation spots only on the lower lip. On endoscopy, polyps were discovered in the proband, her mother, and grandfather. A novel heterozygous mutation (c.521A > C) in exon 4 of STK11 was identified in all four patients, leading to a change from histidine to proline in amino acid 174. The variable site p.H174 was highly conserved in different species on multiple sequence alignment analysis.ConclusionsWe diagnosed a Chinese pedigree with PJS based on clinical features, gastrointestinal endoscopy, and genetic testing results. Our results expanded the spectrum of STK11 variants, which will be helpful for genetic counseling.

Proteome-wide 3D structure prediction provides insights into the ancestral metabolism of ancient archaea and bacteria

Ancestral metabolism has remained controversial due to a lack of evidence beyond sequence-based reconstructions. Although prebiotic chemists have provided hints that metabolism might originate from non-enzymatic protometabolic pathways, gaps between ancestral reconstruction and prebiotic processes mean there is much that is still unknown. Here, we apply proteome-wide 3D structure predictions and comparisons to investigate ancestorial metabolism of ancient bacteria and archaea, to provide information beyond sequence as a bridge to the prebiotic processes. We compare representative bacterial and archaeal strains, which reveal surprisingly similar physiological and metabolic characteristics via microbiological and biophysical experiments. Pairwise comparison of protein structures identify the conserved metabolic modules in bacteria and archaea, despite interference from overly variable sequences. The conserved modules (for example, middle of glycolysis, partial TCA, proton/sulfur respiration, building block biosynthesis) constitute the basic functions that possibly existed in the archaeal-bacterial common ancestor, which are remarkably consistent with the experimentally confirmed protometabolic pathways. These structure-based findings provide a new perspective to reconstructing the ancestral metabolism and understanding its origin, which suggests high-throughput protein 3D structure prediction is a promising approach, deserving broader application in future ancestral exploration.

Deep Local Analysis evaluates protein docking conformations with locally oriented cubes.

MotivationWith the recent advances in protein 3D structure prediction, protein interactions are becoming more central than ever before. Here, we address the problem of determining how proteins interact with one another. More specifically, we investigate the possibility of discriminating near-native protein complex conformations from incorrect ones by exploiting local environments around interfacial residues.ResultsDeep Local Analysis (DLA)-Ranker is a deep learning framework applying 3D convolutions to a set of locally oriented cubes representing the protein interface. It explicitly considers the local geometry of the interfacial residues along with their neighboring atoms and the regions of the interface with different solvent accessibility. We assessed its performance on three docking benchmarks made of half a million acceptable and incorrect conformations. We show that DLA-Ranker successfully identifies near-native conformations from ensembles generated by molecular docking. It surpasses or competes with other deep learning-based scoring functions. We also showcase its usefulness to discover alternative interfaces.Availability and implementation http://gitlab.lcqb.upmc.fr/dla-ranker/DLA-Ranker.git Supplementary information Supplementary data are available at Bioinformatics online.

Best templates outperform homology models in predicting the impact of mutations on protein stability.

Prediction of protein stability change upon mutation (ΔΔG) is crucial for facilitating protein engineering and understanding of protein folding principles. Robust prediction of protein folding free energy change requires the knowledge of protein three-dimensional (3D) structure. In case, protein 3D structure is not available, one can predict the structure from protein sequence; however, the perspectives of ΔΔG predictions for predicted protein structures are unknown. The accuracy of using 3D structures of the best templates for the ΔΔG prediction is also unclear. To investigate these questions, we used a representative set of seven diverse and accurate publicly available tools (FoldX, Eris, Rosetta, DDGun, ACDC-NN, ThermoNet and DynaMut) for stability change prediction combined with AlphaFold or I-Tasser for protein 3D structure prediction. We found that best templates perform consistently better than (or similar to) homology models for all ΔΔG predictors. Our findings imply using the best template structure for the prediction of protein stability change upon mutation if the protein 3D structure is not available. The data are available at https://github.com/ivankovlab/template-vs-model. Supplementary data are available at Bioinformatics online.

A-Prot: protein structure modeling using MSA transformer

BackgroundThe accuracy of protein 3D structure prediction has been dramatically improved with the help of advances in deep learning. In the recent CASP14, Deepmind demonstrated that their new version of AlphaFold (AF) produces highly accurate 3D models almost close to experimental structures. The success of AF shows that the multiple sequence alignment of a sequence contains rich evolutionary information, leading to accurate 3D models. Despite the success of AF, only the prediction code is open, and training a similar model requires a vast amount of computational resources. Thus, developing a lighter prediction model is still necessary.ResultsIn this study, we propose a new protein 3D structure modeling method, A-Prot, using MSA Transformer, one of the state-of-the-art protein language models. An MSA feature tensor and row attention maps are extracted and converted into 2D residue-residue distance and dihedral angle predictions for a given MSA. We demonstrated that A-Prot predicts long-range contacts better than the existing methods. Additionally, we modeled the 3D structures of the free modeling and hard template-based modeling targets of CASP14. The assessment shows that the A-Prot models are more accurate than most top server groups of CASP14.ConclusionThese results imply that A-Prot accurately captures the evolutionary and structural information of proteins with relatively low computational cost. Thus, A-Prot can provide a clue for the development of other protein property prediction methods.

SSpro/ACCpro 6: almost perfect prediction of protein secondary structure and relative solvent accessibility using profiles, deep learning and structural similarity.

Accurately predicting protein secondary structure and relative solvent accessibility is important for the study of protein evolution, structure and an early-stage component of typical protein 3D structure prediction pipelines. We present a new improved version of the SSpro/ACCpro suite of predictors for the prediction of protein secondary structure (in three and eight classes) and relative solvent accessibility. The changes include improved, TensorFlow-trained, deep learning predictors, a richer set of profile features (232 features per residue position) and sequence-only features (71 features per position), a more recent Protein Data Bank (PDB) snapshot for training, better hyperparameter tuning and improvements made to the HOMOLpro module, which leverages structural information from protein segment homologs in the PDB. The new SSpro 6 outperforms the previous version (SSpro 5) by 3-4% in Q3 accuracy and, when used with HOMOLPRO, reaches accuracy in the 95-100% range. The predictors' software, data and web servers are available through the SCRATCH suite of protein structure predictors at http://scratch.proteomics.ics.uci.edu. To maximize comptatibility and ease of use, the deep learning predictors are re-implemented as pure Python/numpy code without TensorFlow dependency. Supplementary data are available at Bioinformatics online.

The ABA receptor gene MdPYL9 confers tolerance to drought stress in transgenic apple (Malus domestica)

Abscisic acid (ABA) is an important phytohormone for plant growth, development, and the response to various types of abiotic stress. Pyrabactin Resistance 1 (PYR1)/PYR1-Like (PYL)/Regulatory Component of ABA Receptor (RCAR) family proteins (PYLs) are ABA receptors that have been studied in many plant species. However, no detailed functional studies of these genes in regulating apple drought tolerance have been reported. Here, we isolated a homolog of the ABA receptor PYL9 in Arabidopsis, MdPYL9, from apple (Malus domestica) and analyzed its role in regulating drought tolerance in transgenic apple plants. Gene structure and protein sequence analysis showed that MdPYL9 shares a high similarity with AtPYL9. Protein sequence comparison and 3D structure prediction identified several conserved motifs and residues that are important for ABA binding in MdPYL9, including the helix-grip structure and the gate and latch loops. Expression analysis showed that MdPYL9 was significantly induced by drought treatment. Overexpression of MdPYL9 conferred enhanced tolerance to drought stress in transgenic apple plants, including a higher photosynthetic rate and relative water content and lower MDA content and electrolyte leakage compared with wild-type plants. Lower accumulation of reactive oxygen species (ROS) and higher activity of antioxidant enzymes suggest that MdPYL9 overexpression could promote ROS scavenging and thus reduce oxidative damage in transgenic apple plants under drought stress. Yeast two-hybrid assays indicated that there were both ABA-dependent and ABA-independent interactions between MdPYL9 and the clade A subgroup members of the PP2C family (PP2CAs) in apple. Expression analysis identifed two PP2CA genes, MdHAI1 and MdHAI2, that were highly expressed under drought treatment by not expressed under normal conditions. These results provide new insight that could be used to aid future studies examining the function of MdPYLs and the mechanism by which they regulate apple drought tolerance.

Highly Conserved C-Terminal Region of Indian Hedgehog N-Fragment Contributes to Its Auto-Processing and Multimer Formation.

Hedgehog (HH) is a highly conserved secretory signalling protein family mainly involved in embryonic development, homeostasis, and tumorigenesis. HH is generally synthesised as a precursor, which subsequently undergoes autoproteolytic cleavage to generate an amino-terminal fragment (HH-N), mediating signalling, and a carboxyl-terminal fragment (HH-C), catalysing the auto-processing reaction. The N-terminal region of HH-N is required for HH multimer formation to promote signal transduction, whilst the functions of the C-terminal region of HH-N remain ambiguous. This study focused on Indian Hedgehog (IHH), a member of the HH family, to explore the functions of the C-terminal region of the amino-terminal fragment of IHH (IHH-N) via protein truncation, cell-based assays, and 3D structure prediction. The results revealed that three amino acids, including S195, A196, and A197, were crucial for the multimer formation by inserting the mutual binding of IHH-N proteins. K191, S192, E193, and H194 had an extremely remarkable effect on IHH self-cleavage. In addition, A198, K199, and T200 evidently affected the stability of IHH-N. This work suggested that the C-terminus of IHH-N played an important role in the physiological function of IHH at multiple levels, thus deepening the understanding of HH biochemical properties.

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

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

Predicting Local Protein 3D Structures Using Clustering Deep Recurrent Neural Network.

Since protein 3D structure prediction is very important for biochemical study and drug design, researchers have developed many machine learning algorithms to predict protein 3D structures using the sequence information only. Understanding the sequence-to-structure relationship is key for the successful structure prediction. Previous approaches including the single shallow learning model, the single deep learning model and clustering algorithms all have disadvantages to understand precise sequence-to-structure relationship. In order to further improve the performance of the local protein structure prediction, a novel deep learning model called Clustering Recurrent Neural Network (CRNN) is proposed. In this model, the whole protein dataset is divided into multiple cluster subtrees. A RNN is trained for each cluster in the subtrees so that each RNN can be used to learn the computationally simpler local sequence-to-structure relationship instead of attempting to capture the global sequence-to-structure relationship. After learning the local sequence-to-structure relationship using RNN, CRNN is designed to predict distance matrices, torsion angles and secondary structures for backbone α-carbon atoms of protein sequence segments. Our experimental analysis indicates that 3D structure prediction accuracy is comparable or better than other state-of-art approaches.

Isolation and Characterization of C-C Chemokine Ligand 7 (CCL7) in Cynomolgus Macaques

Cynomolgus macaques (Macaca fascicularis) are an established animal model of asthma, which exhibit different responses to allergen exposure that are clinically relevant. The chemokine ligand gene (CCL7) encodes Monocyte Chemotactic Protein-3, which has an important role in asthma pathogenesis. While CCL7 polymorphism in humans is associated with asthma phenotype, very little is known about CCL7 in nonhuman primate models of respiratory disease. The objective of this study was to isolate and characterize CCL7 gene in cynomolgus macaques of Indonesian origin. In this study, we used sequencing and bioinformatics technique for gene isolation, characterization, and protein 3D structure prediction. We isolated a 2253 base-pair (bp) sequence of CCL7 in cynomolgus macaques, which exhibited 95% similarity in coding sequence to human CCL7. The amino acid sequence was more closely clustered with human CCL7 than with that of rodents. Importantly, the predictive protein structure of CCL7 was similar to that in humans. These similarities in CCL7 suggests the potential of cynomolgus macaque as a translational model to study asthma, particularly in the context of genetics and role of chemokines such as CCL7.

Isolation and Characterization of C-C Chemokine Ligand 7 (CCL7) in Cynomolgus Macaques

Cynomolgus macaques (Macaca fascicularis) are an established animal model of asthma, which exhibit different responses to allergen exposure that are clinically relevant. The chemokine ligand gene (CCL7) encodes Monocyte Chemotactic Protein-3, which has an important role in asthma pathogenesis. While CCL7 polymorphism in humans is associated with asthma phenotype, very little is known about CCL7 in nonhuman primate models of respiratory disease. The objective of this study was to isolate and characterize CCL7 gene in cynomolgus macaques of Indonesian origin. In this study, we used sequencing and bioinformatics technique for gene isolation, characterization, and protein 3D structure prediction. We isolated a 2253 base-pair (bp) sequence of CCL7 in cynomolgus macaques, which exhibited 95% similarity in coding sequence to human CCL7. The amino acid sequence was more closely clustered with human CCL7 than with that of rodents. Importantly, the predictive protein structure of CCL7 was similar to that in humans. These similarities in CCL7 suggests the potential of cynomolgus macaque as a translational model to study asthma, particularly in the context of genetics and role of chemokines such as CCL7.

ResPRE: high-accuracy protein contact prediction by coupling precision matrix with deep residual neural networks.

Contact-map of a protein sequence dictates the global topology of structural fold. Accurate prediction of the contact-map is thus essential to protein 3D structure prediction, which is particularly useful for the protein sequences that do not have close homology templates in the Protein Data Bank. We developed a new method, ResPRE, to predict residue-level protein contacts using inverse covariance matrix (or precision matrix) of multiple sequence alignments (MSAs) through deep residual convolutional neural network training. The approach was tested on a set of 158 non-homologous proteins collected from the CASP experiments and achieved an average accuracy of 50.6% in the top-L long-range contact prediction with L being the sequence length, which is 11.7% higher than the best of other state-of-the-art approaches ranging from coevolution coupling analysis to deep neural network training. Detailed data analyses show that the major advantage of ResPRE lies at the utilization of precision matrix that helps rule out transitional noises of contact-maps compared with the previously used covariance matrix. Meanwhile, the residual network with parallel shortcut layer connections increases the learning ability of deep neural network training. It was also found that appropriate collection of MSAs can further improve the accuracy of final contact-map predictions. The standalone package and online server of ResPRE are made freely available, which should bring important impact on protein structure and function modeling studies in particular for the distant- and non-homology protein targets. https://zhanglab.ccmb.med.umich.edu/ResPRE and https://github.com/leeyang/ResPRE. Supplementary data are available at Bioinformatics online.

PSSRDBModel- Protein 3D structure prediction server based on the secondary structure informations

Abstract The 3D structure of a protein is highly depended on its sequence. In the absence of an experimentally determined structure, comparative or homology modeling can provide a useful 3D model for a protein that is related to at least one known protein structure (template). Comparative modeling predicts the 3D structure of a given protein sequence (target) primarily based on its alignment with one or more proteins of known structure (templates). However, the sequence similarities below 25% (twilight zone) are not considered for modeling. In this scenario, we have emphasized approach to predict the 3D structure by considering the similarity of secondary structural elements and their consensus. We have developed a database with Protein secondary structure representative database (PSSRD) that holds the secondary structure information. In this work we have designed a PSSRDBModel server that can model the proteins by selecting the template by considering the templates secondary structure element information and Medical subject heading (MeSH) terms. The modeling server at PSSRDB is based on the sequence homology based on the target-template alignment and MeSH terms and uses MODELLER algorithm.

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

PredMP is the first web service, to our knowledge, that aims at de novo prediction of the membrane protein (MP) 3D structure followed by the embedding of the MP into the lipid bilayer for visualization. Our approach is based on a high-throughput Deep Transfer Learning (DTL) method that first predicts MP contacts by learning from non-MPs and then predicts the 3D model of the MP using the predicted contacts as distance restraints. This algorithm is derived from our previous Deep Learning (DL) method originally developed for soluble protein contact prediction, which has been officially ranked No. 1 in CASP12. The DTL framework in our approach overcomes the challenge that there are only a limited number of solved MP structures for training the deep learning model. There are three modules in the PredMP server: (i) The DTL framework followed by the contact-assisted folding protocol has already been implemented in RaptorX-Contact, which serves as the key module for 3D model generation; (ii) The 1D annotation module, implemented in RaptorX-Property, is used to predict the secondary structure and disordered regions; and (iii) the visualization module to display the predicted MPs embedded in the lipid bilayer guided by the predicted transmembrane topology. Tested on 510 non-redundant MPs, our server predicts correct folds for ∼290 MPs, which significantly outperforms existing methods. Tested on a blind and live benchmark CAMEO from September 2016 to January 2018, PredMP can successfully model all 10 MPs belonging to the hard category. PredMP is freely accessed on the web at http://www.predmp.com. Supplementary data are available at Bioinformatics online.

Protein threading using residue co-variation and deep learning.

MotivationTemplate-based modeling, including homology modeling and protein threading, is a popular method for protein 3D structure prediction. However, alignment generation and template selection for protein sequences without close templates remain very challenging.ResultsWe present a new method called DeepThreader to improve protein threading, including both alignment generation and template selection, by making use of deep learning (DL) and residue co-variation information. Our method first employs DL to predict inter-residue distance distribution from residue co-variation and sequential information (e.g. sequence profile and predicted secondary structure), and then builds sequence-template alignment by integrating predicted distance information and sequential features through an ADMM algorithm. Experimental results suggest that predicted inter-residue distance is helpful to both protein alignment and template selection especially for protein sequences without very close templates, and that our method outperforms currently popular homology modeling method HHpred and threading method CNFpred by a large margin and greatly outperforms the latest contact-assisted protein threading method EigenTHREADER.Availability and implementation http://raptorx.uchicago.edu/ Supplementary information Supplementary data are available at Bioinformatics online.