Development server - no detailed description available

Abstract

The Naive AlphaFoldDB 100 method is a baseline that fetches models with 100% sequence identity to the target sequences from the AlphaFold Protein Structure Database at EMBL-EBI Alignents are retrieved with the FASTA REST API. Models corresponding to the top hits are cut and renumbered according to the target sequence. No remodeling is performed.

Tunyasuvunakool, K., Adler, J., Wu, Z. et al. Highly accurate protein structure prediction for the human proteome. Nature (2021). DOI: 10.1038/s41586-021-03828-1

Madeira F., Pearce M., Tivey A. R. N. et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Research (2022) gkac240. DOI: 10.1093/nar/gkac240

Abstract

Protein structure prediction plays an important role in drug discovery. In order to achieve robust performance, PaFold first obtains massive co-evolutional based information from raw sequence via huge protein sequence and template databases. Several layers of deep neural networks are then applied to predict the coarse-grained motifs and decoys. These, together with the previous template based information, can accurately guide the construction of 3D protein structure. The structure then undergoes several rounds of refinement based on energy minimization to achieve a more precise outcome.

Development server - no detailed description available

Development server - no detailed description available

Abstract

To uncover the mystery of the sequence to structure to function relationship is of great significance to the humankind for better understanding our lives. In order to achieve this dream, HeliXonAI, an integrated AI-enabled drug design platform, takes the giant leap to reach the state-of-the-art performance on protein structure prediction leveraging both informative sequence evolutionary features and modern geometric machine learning algorithms.

Abstract

Following PaFold, PaFold_v2 further accelerates and improves the overall prediction performance by means of Transformer and self-attention mechanism. We also develop a novel selection algorithm based on diversity and distribution to achieve more robust intermediate data. In general, it can still give solid results even without template information and limited input data.

Abstract

The successful application of deep learning has promoted a breakthrough progress in protein structure prediction. Generally, deep learning is independently used to predict the contact/distance between residues or to evaluate the accuracy of the model. RocketX integrates model evaluation into geometric constraint prediction and structural modeling to build a feedback mechanism to achieve closed-loop structural optimization. In RocketX, two different deep residual neural networks are designed to predict the inter-residue geometric constraints and evaluate the quality of the folded model, respectively. The predicted geometric constraints are used to guide the structural model folding. The model evaluation network is used to estimate the quality of the folded model, and the results are fed back to the geometric constraint prediction network to re-predict the geometric constraints and fold the new structural model. The final structural model is generated through three closed-loop iterative optimizations.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

OpenComplex is an open-source platform for developing protein and RNA complex models.

The code is available on GitHub.

Abstract

The successful application of deep learning has promoted a breakthrough progress in protein structure prediction, such as AlphaFold2. However, the full-chain modelling appears to be lower on average accuracy than that for the constituent domains and requires higher demand on computing hardware, indicating the performance of full-chain modelling still needs to be improved. In SADA Server, we first used protein domain boundary prediction method DomBpred [1] to predict domain boundary of input sequence. Then, AlphaFold2 is used to predict the structure of each domain sequence. Finally, a protein domain assembly method [2] is used to assemble the predicted domain model to generate the full-length model.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

The Robetta server (http://robetta.bakerlab.org) provides automated tools for protein structure prediction and analysis. For structure prediction, sequences submitted to the server are parsed into putative domains and structural models are generated using either comparative modeling or de novo structure prediction methods. If a confident match to a protein of known structure is found using BLAST, PSI-BLAST, FFAS03 or 3D-Jury, it is used as a template for comparative modeling. If no match is found, structure predictions are made using the de novo Rosetta fragment insertion method. Experimental nuclear magnetic resonance (NMR) constraints data can also be submitted with a query sequence for RosettaNMR de novo structure determination. Other current capabilities include the prediction of the effects of mutations on protein–protein interactions using computational interface alanine scanning. The Rosetta protein design and protein–protein docking methodologies will soon be available through the server as well.

• Kim DE, Chivian D, Baker D. Protein structure prediction and analysis using the Robetta server. (2004) Nucleic Acids Res. 32(Web Server issue):W526-31. (DOI: 10.1093/nar/gkh468)
• Raman S, Vernon R, et al. Structure prediction for CASP8 with all-atom refinement using Rosetta. (2009) Proteins 77 Suppl 9:89-99.i (DOI: 10.1002/prot.22540)
• Song Y, DiMaio F, et al. High resolution comparative modeling with RosettaCM. (2013) Structure 21(10):1735-42. (DOI: 10.1016/j.str.2013.08.005)
• Yang, J, Anishchenko I, et al. Improved Protein Structure Prediction Using Predicted Interresidue Orientations. (2020) Proceedings of the National Academy of Sciences of the United States of America 117 (3):1496–1503. (DOI: 10.1073/pnas.1914677117)

Abstract

DeepAssembly uses a multi-domain assembly approach to predict full-length protein structures. In DeepAssembly, we first constructed a deep learning model, AffineNet, which is specially used to predict the affine transformations of inter-domain residues. Then, an energy function called Atomic Coordinate Deviation potential was designed according to the predicted affine transformations. Finally, in domain assembly module, the energy function was optimized by population-based optimization method, and the single-domain structures was assembled, so as to obtain the full-length model.

Abstract

The successful application of deep learning has promoted a breakthrough progress in protein structure prediction, such as AlphaFold2, RoseTTAFold, trRosetta. The MultiDFold server is an attempt to fuse multiple information for structure prediction based on the multi-objective optimization method.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

Advanced protein structure prediction models are sensitive to the co-evolution inputs. AirFold, an automated participating server, aims to improve efficiency and performance of protein structure prediction module by exploring better representation of sequence evolutionary features.

Development server - no detailed description available

Abstract

Modelling the 3D structures of proteins can often be enhanced if more than one fold template is used during the modelling process. However, in many cases, this may also result in poorer model quality for a given target or alignment method. There is a need for modelling protocols that can both consistently and significantly improve 3D models and provide an indication of when models might not benefit from the use of multiple target-template alignments. Here, we investigate the use of both global and local model quality prediction scores produced by ModFOLDclust2, to improve the selection of target-template alignments for the construction of multiple-template models. Additionally, we evaluate clustering the resulting population of multi- and single-template models for the improvement of our IntFOLD-TS tertiary structure prediction method.

Buenavista, M. T., Roche, D. B. & McGuffin, L. J. (2012) Improvement of 3D protein models using multiple templates guided by single-template model quality assessment.

Bioinformatics, 28, 1851-1857.

http://dx.doi.org/10.1093/bioinformatics/bts292

Abstract

PAthreader is a high-precision remote template detection method by threading the PDB and AFDB libraries. AlphaFold2 is improved by PAthreader by providing better template alignment.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

The SAIS-Fold server strives to seamlessly integrate a wide range of information sources for accurate structure prediction. It achieves this by synergistically combining an expertise-driven approach with the Protein Language Model (PLM) method.

Abstract

DeepSHfold is a structure prediction method based on various MSA sampling. In particular, a search for structural homologous sequences based on protein language models is proposed. And for complexes, a pair-MSA strategy based on sequence structure homology scores was designed to enhance sampling.

Development server - no detailed description available

Development server - no detailed description available

Abstract

Improvements in comparative protein structure modeling for the remote target-template sequence similarity cases are possible through the optimal combination of multiple template structures and by improving the quality of target-template alignment. Recently developed MMM and M4T methods were designed to address these problems. Here we describe new developments in both the alignment generation and the template selection parts of the modeling algorithms. We set up a new scoring function in MMM to deliver more accurate target-template alignments. This was achieved by developing and incorporating into the composite scoring function a novel statistical pairwise potential that combines local and non-local terms. The non-local term of the statistical potential utilizes a shuffled reference state definition that helped to eliminate most of the false positive signal from the background distribution of pairwise contacts. The accuracy of the scoring function was further increased by using BLOSUM mutation table scores.

Rykunov D, Steinberger E, Madrid-Aliste CJ, Fiser A Improved scoring function for comparative modeling using the M4T method.

J Struct Funct Genomics (2009) 10(1) : 95-9.

http://dx.doi.org/10.1007/s10969-008-9044-9

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

Determining the structure and function of a novel protein is a cornerstone of many aspects of modern biology. Over the past decades, a number of computational tools for structure prediction have been developed. It is critical that the biological community is aware of such tools and is able to interpret their results in an informed way. This protocol provides a guide to interpreting the output of structure prediction servers in general and one such tool in particular, the protein homology/analogy recognition engine (Phyre). New profile-profile matching algorithms have improved structure prediction considerably in recent years. Although the performance of Phyre is typical of many structure prediction systems using such algorithms, all these systems can reliably detect up to twice as many remote homologies as standard sequence-profile searching. Phyre is widely used by the biological community, with >150 submissions per day, and provides a simple interface to results. Phyre takes 30 min to predict the structure of a 250-residue protein.

Kelley LA, Sternberg MJ. Protein structure prediction on the Web: a case study using the Phyre server.

Nat Protoc. 2009;4(3):363-71

http://dx.doi.org/10.1038/nprot.2009.2

Development server - no detailed description available

Abstract

DeepMind presented remarkably accurate protein structure predictions at the CASP14 conference. We explored network architectures incorporating related ideas and obtained the best performance with a three-track attention-based network in which information at the 1D sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The 3-track network produces high accuracy structure predictions, enables rapid solution of challenging X-ray crystallography and cryo-EM structure modeling problems, and provides insights into the functions of proteins of currently unknown structure.

Abstract

Protein structure homology modelling has become a routine technique to generate 3D models for proteins when experimental structures are not available. Fully automated servers such as SWISS-MODEL with user-friendly web interfaces generate reliable models without the need for complex software packages or downloading large databases. Here, we describe the latest version of the SWISS-MODEL expert system for protein structure modelling. The SWISS-MODEL template library provides annotation of quaternary structure and essential ligands and co-factors to allow for building of complete structural models, including their oligomeric structure. The improved SWISS-MODEL pipeline makes extensive use of model quality estimation for selection of the most suitable templates and provides estimates of the expected accuracy of the resulting models. The accuracy of the models generated by SWISS-MODEL is continuously evaluated by the CAMEO system. The new web site allows users to interactively search for templates, cluster them by sequence similarity, structurally compare alternative templates and select the ones to be used for model building. In cases where multiple alternative template structures are available for a protein of interest, a user-guided template selection step allows building models in different functional states. SWISS-MODEL is available at http://swissmodel.expasy.org/.

Biasini, M. et.al. Nucleic Acids Research; (1 July 2014) 42 (W1): W252-W258

Development server - no detailed description available

Abstract

A key challenge of modern biology is to uncover the functional role of the protein entities that compose cellular proteomes. To this end, the availability of reliable three-dimensional atomic models of proteins is often crucial. This protocol presents a community-wide web-based method using RaptorX (http://raptorx.uchicago.edu/) for protein secondary structure prediction, template-based tertiary structure modeling, alignment quality assessment and sophisticated probabilistic alignment sampling. RaptorX distinguishes itself from other servers by the quality of the alignment between a target sequence and one or multiple distantly related template proteins (especially those with sparse sequence profiles) and by a novel nonlinear scoring function and a probabilistic-consistency algorithm. Consequently, RaptorX delivers high-quality structural models for many targets with only remote templates. At present, it takes RaptorX ∼35 min to finish processing a sequence of 200 amino acids.

Källberg M, Wang H, Wang S, Peng J, Wang Z, Lu H, Xu J. Template-based protein structure modeling using the RaptorX web server.

Nature Protocols 7, 1511–1522, 2012.

http://dx.doi.org/10.1038/nprot.2012.085

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

The Princeton_TEMPLATE (ProTEin geoMetry Prediction using simuLAtions and supporT vEctor machines) webserver is an automated template-assisted structure prediction method that produces 5 model predictions for targets with sufficient homology to a template. First, the method does single or multiple domain alignments to template structures contained in our newly constructed template library. Next, sequence alignments are generated using the SPARKS-X Fold Recognition Software developed by the Yaoqi Zhou Lab. Next, the method generates many local minima based on the constraints generated using torsion-angle dynamics in Cyana. Each of these local minima are next relaxed using Rosetta Fast Relax with constraints on the coordinates to remain near the start in order to repack the side-chains and to make movements that will enhance the number of hyrogen bonds. This ensemble of structures is refined using components of the Princeton_TIGRESS refinement protocol. Namely, the structures are passed through a support vector machines based function evaluation that is capable of identifying models that are improved relative to the naive lowest-energy model. The model selected by the SVM is then refined via all atom molecular dynamics simulatons in CHARMM using the FACTS implicit solvent model. The final prediction and refined structure is sent to the user. The methodology has been benchmarked on all CASP10 targets.

Khoury, G. A. et.al. Proteins: Structure, Function, and Bioinformatics 2014, 82 (5), 794-814.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

In recent years, development of a single-method fold-recognition server lags behind consensus and multiple template techniques. However, a good consensus prediction relies on the accuracy of individual methods. This article reports our efforts to further improve a single-method fold recognition technique called SPARKS by changing the alignment scoring function and incorporating the SPINE-X techniques that make improved prediction of secondary structure, backbone torsion angle and solvent accessible surface area.

Yang Y, Faraggi E, Zhao H, Zhou Y. Improving protein fold recognition and template-based modeling by employing probabilistic-based matching between predicted one-dimensional structural properties of query and corresponding native properties of templates.Bioinformatics.

2011 Aug 1;27(15):2076-82.

http://dx.doi.org/10.1093/bioinformatics/btr350

Abstract

The NaiveBlits method servers as baseline by relying on the first hit found by HHblits as template and predicting the protein structure by calling MODELLER.

Abstract

RBO Aleph is a novel protein structure prediction web server for template-based modeling, protein contact prediction and ab initio structure prediction. The server has a strong emphasis on modeling difficult protein targets for which templates cannot be detected. RBO Aleph's unique features are (i) the use of combined evolutionary and physicochemical information to perform residue–residue contact prediction and (ii) leveraging this contact information effectively in conformational space search. RBO Aleph emerged as one of the leading approaches to ab initio protein structure prediction and contact prediction during the most recent Critical Assessment of Protein Structure Prediction experiment (CASP11, 2014). In addition to RBO Aleph's main focus on ab initio modeling, the server also provides state-of-the-art template-based modeling services. Based on template availability, RBO Aleph switches automatically between template-based modeling and ab initio prediction based on the target protein sequence, facilitating use especially for non-expert users. The RBO Aleph web server offers a range of tools for visualization and data analysis, such as the visualization of predicted models, predicted contacts and the estimated prediction error along the model's backbone. The server is accessible at http://compbio.robotics.tu-berlin.de/rbo_aleph/.

Mabrouk, M., Putz, I., Werner, T., Schneider, M., Neeb, M., Bartels, P. and Brock, O., 2015. RBO Aleph: leveraging novel information sources for protein structure prediction. Nucleic acids research, 43(W1), pp.W343-W348.

dx.doi.org/10.1093/nar/gkv357

Abstract

Modelling the 3D structures of proteins can often be enhanced if more than one fold template is used during the modelling process. However, in many cases, this may also result in poorer model quality for a given target or alignment method. There is a need for modelling protocols that can both consistently and significantly improve 3D models and provide an indication of when models might not benefit from the use of multiple target-template alignments. Here, we investigate the use of both global and local model quality prediction scores produced by ModFOLDclust3, to improve the selection of target-template alignments for the construction of multiple-template models. Additionally, we evaluate clustering the resulting population of multi- and single-template models for the improvement of our IntFOLD-TS tertiary structure prediction method.

McGuffin, L.J., Atkins, J., Salehe, B.R., Shuid, A.N. & Roche, D.B. (2015) IntFOLD: an integrated server for modelling protein structures and functions from amino acid sequences. Nucleic Acids Research, 43, W169-73.

Development server - no detailed description available

Development server - no detailed description available

Abstract

The NaiveBlast method servers as baseline by relying on the first hit found by BLAST as template and predicting the protein structure by calling MODELLER.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

Automated protein structure prediction is becoming a mainstream tool for biological research. This has been fueled by steady improvements of publicly available automated servers over the last decade, in particular their ability to build good homology models for an increasing number of targets by reliably detecting and aligning more and more remotely homologous templates. Here, we describe the three fully automated versions of the HHpred server that participated in the community-wide blind protein structure prediction competition CASP8. What makes HHpred unique is the combination of usability, short response times (typically under 15 min) and a model accuracy that is competitive with those of the best servers in CASP8.

Hildebrand A., Remmert M., Biegert A., and Söding J. Fast and accurate automatic structure prediction with HHpred.

Proteins. 77 Suppl 9:128-132 (2009).

dx.doi.org/10.1002/prot.22499

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

For our automated predictions we developed the IntFOLD4-TS protocol, which integrates the ModFOLD6_rank method for scoring the multiple-template models that were generated using a number of alternative sequence-structure alignments. Overall, our selection of top models and Accuracy Self Estimate (ASE) scores using ModFOLD6_rank was an improvement on our previous approaches. Our IntFOLD4-TS method was developed with the aim of identifying, and then attempting to fix, the local errors in an initial pool of single template models via iterative multi-template modeling. The method attempts to exploit our previous CASP successes in accurately predicting local errors in our models by taking the global and local per-residue errors into consideration during the multiple template selection stage.The pipeline can be broken down into two major stages: (1) single template modeling with ASE scoring and (2) QA guided multiple template modeling with ASE scoring.

McGuffin, L.J., Shuid, A.M., Kempster, R., Maghrabi, A.H.A., Nealon J.O., Salehe, B.R., Atkins, J.D. & Roche, D.B. (2017) Accurate Template Based Modelling in CASP12 using the IntFOLD4-TS, ModFOLD6 and ReFOLD methods. Proteins: Structure, Function, and Bioinformatics, 86 Suppl 1, 335-344. McGuffin, L.J., Atkins, J., Salehe, B.R., Shuid, A.N. & Roche, D.B. (2015) IntFOLD: an integrated server for modelling protein structures and functions from amino acid sequences. Nucleic Acids Research, 43, W169-73.

Development server - no detailed description available

Abstract

The IntFOLD-TS method was developed according to the guiding principle that the model quality assessment (QA) would be the most critical stage for our template-based modeling pipeline. Thus, the IntFOLD-TS method firstly generates numerous alternate models, using in-house versions of several different sequence-structure alignment methods, which are then ranked in terms of global quality using our top performing QA method-ModFOLDclust2. In addition to the predicted global quality scores, the predictions of local errors are also provided in the resulting coordinate files, using scores that represent the predicted deviation of each residue in the model from the equivalent residue in the native structure. The IntFOLD-TS method was found to generate high quality 3D models for many of the CASP9 targets, whilst also providing highly accurate predictions of their per-residue errors. This important information may help to make the 3D models that are produced by the IntFOLD-TS method more useful for guiding future experimental work

Roche, D. B., Buenavista, M. T., Tetchner, S. J. & McGuffin, L. J.

(2011) The IntFOLD server: an integrated web resource for protein fold recognition, 3D model quality assessment, intrinsic disorder prediction, domain prediction and ligand binding site prediction. Nucleic Acids Res., 39, W171-6.

http://www.ncbi.nlm.nih.gov/pubmed/21459847

Development server - no detailed description available

Abstract

The development of automated servers to predict the three-dimensional structure of proteins has seen much progress over the years. These servers make calculations simpler, but largely exclude users from the process. In this study, we present the PRotein Interactive MOdeling (PRIMO) pipeline for homology modeling of protein monomers. The pipeline eases the multi-step modeling process, and reduces the workload required by the user, while still allowing engagement from the user during every step. Default parameters are given for each step, which can either be modified or supplemented with additional external input. PRIMO has been designed for users of varying levels of experience with homology modeling. The pipeline incorporates a user-friendly interface that makes it easy to alter parameters used during modeling. During each stage of the modeling process, the site provides suggestions for novice users to improve the quality of their models. PRIMO provides functionality that allows users to also model ligands and ions in complex with their protein targets. Herein, we assess the accuracy of the fully automated capabilities of the server, including a comparative analysis of the available alignment programs, as well as of the refinement levels used during modeling. The tests presented here demonstrate the reliability of the PRIMO server when producing a large number of protein models. While PRIMO does focus on user involvement in the homology modeling process, the results indicate that in the presence of suitable templates, good quality models can be produced even without user intervention. This gives an idea of the base level accuracy of PRIMO, which users can improve upon by adjusting parameters in their modeling runs. The accuracy of PRIMO’s automated scripts is being continuously evaluated by the CAMEO (Continuous Automated Model EvaluatiOn) project.

Hatherley R, Brown DK, Glenister M, Tastan Bishop Ö (2016) PRIMO: An Interactive Homology Modeling Pipeline. PLOS ONE 11(11): e0166698.

https://doi.org/10.1371/journal.pone.0166698

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

The tFold server integrates three innovative techniques for high-accuracy protein structure prediction. We adopt the “multi-source fusion” strategy to fully exploit co-evolution patterns embedded in multiple groups of MSA data. An ultra-deep criss-cross attention residual network is developed to accurately predict both inter-residue distance and orientation relationships. We further develop the “template-based free modeling - TBFM” framework to combine both TBM-based tertiary structure information and FM-based distance and orientation predictions to generate high-quality structure predictions.

Han, Y., Zhuang, Q., Sun, B. et al. Crystal structure of steroid reductase SRD5A reveals conserved steroid reduction mechanism. Nat Commun 12, 449 (2021). DOI: 10.1038/s41467-020-20675-2.

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Development server - no detailed description available

Abstract

The BestSingleStructuralTemplate method serves as "post-diction" baseline, representing an upper limit for single template models. The best templates are discovered by structural superposition of the target reference structures with all PDB structures using TM‐align. The top 20 of the obtained structural alignments serve as input for the subsequent template‐based modeling. Modeling is performed with SWISS‐MODEL's modeling engine ProMod3. Termini beyond the region covered by the template structure are modeled by a low‐complexity Monte Carlo sampling approach. The final models are ranked by lDDT, and the top scoring model is selected for that particular target.

Haas, J, Gumienny, R, Barbato, A, et al. Introducing “best single template” models as reference baseline for the Continuous Automated Model Evaluation (CAMEO). Proteins. 2019; 87: 1378–1387.

https://doi.org/10.1002/prot.25815

Abstract

The Naive AlphaFoldDB 90 method is a baseline that fetches models with at least 90% sequence identity to the target sequences from the AlphaFold Protein Structure Database at EMBL-EBI. Alignents are retrieved with the FASTA REST API. Models corresponding to the top hits are downloaded and minimal modelling required to map the AlphaFold model to the target sequence is performed with ProMod3 using its default modelling pipeline.

Tunyasuvunakool, K., Adler, J., Wu, Z. et al. Highly accurate protein structure prediction for the human proteome. Nature (2021). DOI: 10.1038/s41586-021-03828-1

Studer, G., Tauriello, G., Bienert S. et al. ProMod3—A versatile homology modelling toolbox. PLOS Computational Biology (2021) 17(1): e1008667. DOI: 10.1371/journal.pcbi.1008667

Madeira F., Pearce M., Tivey A. R. N. et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Research (2022) gkac240. DOI: 10.1093/nar/gkac240

Development server - no detailed description available