OpenStructure Actions API · June 2025
OpenStructure, the open-source molecular modelling environment that forms the basis of SWISS-MODEL and the Structure Assessment Web Server has recently introduced a benchmarking suite, available as a set of 'actions'. Currently the method is published as a preprint.
Two of these actions focus on comparing theoretical models of macromolecular complexes with their corresponding reference structures, providing direct access to evaluation scores:
- compare-structures: Compares structures consisting of any number of polymer entities.
- compare-ligand-structures: Compares interactions between polymer entities and non-polymer entities, such as small-molecule ligands.
These OpenStructure Actions can be used in several ways: via Docker, Singularity, or Conda; compiled from source; or accessed directly through the OpenStructure Actions API.
Structure Assessment web server publication · April 2024
The Structure Assessment web server has been published in the 2024 Nucleic Acids Research Web Server issue.
38634802
10.1093/nar/gkae270It is a one-stop shop for interactive evaluation and benchmarking of structural models of macromolecular complexes including proteins and nucleic acids. A user-friendly web dashboard links sequence with structure information and results from a variety of state-of-the-art tools, which facilitates the visual exploration and evaluation of structure models. The dashboard integrates stereochemistry information, secondary structure information, global and local model quality assessment of the tertiary structure of comparative protein models, as well as prediction of membrane location. In addition, a benchmarking mode is available where a model can be compared to a reference structure, providing easy access to scores that have been used in recent CASP experiments and CAMEO.
AlphaFoldDB structures included as homology modelling templates · April 2023
With the structural coverage of known proteins in UniProt greatly expanded by the AlphaFold Protein Structure Database (AFDB), these computed structures are now included alongside the experimental structure based templates from the SWISS-MODEL Template Library in the SWISS-MODEL homology modelling pipeline.
If the target sequence can be found by a fast k-mer search of the AFDB at over 70% sequence identity, it will be added to the template results, assigned a GMQE score for ranking, and it will be considered for selection in the automated modelling mode. Currently, this is only used for projects with a single target sequence since AFDB only includes monomers.
Assess structure against reference · April 2023
lDDT, QS-Score, TM-Score, RMSD and DockQ are available with the Structure Assessment service.
lDDT has been greatly enhanced to allow comparison of multiple chains as well as nucleic acids. This implementation is available in OpenStructure and supersedes the original lDDT implementation which has been deprecated.
When multimers are assessed, the interface similarity is measured using QS-score which quantifies the similarity between interfaces as a function of shared interface contacts.
OpenStructure now has an implementation of DockQ-wave, a variant of the DockQ score, which aims to provide a continuous protein-protein docking model quality measure.
Furthermore, TM-score, as computed with US-align, assesses the topological similarity of macromolecular complexes including RNA.
You can find out more about it in our help page.
Model Annotations · July 2022
We have added a simple way to annotate your models. Try it out for the results of your modelling or structure assessment projects by clicking on the pen icon below the 3D view of the model.
You will be able to identify regions by providing sequence motifs, chain names or residue numbers. The selected regions can then be coloured and labelled according to your needs. You can find out more about it in our help page.
Modelling API · March 2022
There have been many email requests to our helpdesk about the availability of an API for SWISS-MODEL, finally you are now able to send your own GET and POST requests to our API service!
Start modelling jobs from the command line using curl, your programming language of choice or try out the available endpoints with either Swagger UI or Core API.
We have prepared some minimal Python code to get you started on our help page.
TBvar3D · February 2022
Check out the latest addition to the SWISS-MODEL workspace: TBvar3D, a web server for the interpretation of antibiotic resistance variants of Mycobacterium tuberculosis on protein structures.
A sleek interface integrates the WHO M. tuberculosis (MTB) resistance variant catalog with manually curated protein structures. Antibiotic resistance variants can now directly be explored in a structural context and users can submit custom MTB variants to be analyzed and visualized by TBvar3D.
Sequence Feature Display · July 2021
The display of structures and sequence annotations in SWISS-MODEL Repository was upgraded to improve the navigation. This allows multiple linear tracks of user annotations to be compared simultaneously.
It is now possible to create a link to display all your annotations on a particular entry in Repository making it easier to share with others:
Model Quality with QMEANDisCo July 2021
The QMEANDisCo global score is now the primary global model quality predictor throughout all SWISS-MODEL services. An intuitive color scheme informs about model quality in a glance.
Quality measures based on QMEAN4 Z-scores are still available but considered deprecated.
As a consequence, we also adapted the quality filter for the SWISS-MODEL Repository model import. A QMEANDisCo global score threshold of 0.5 significantly reduces the likelihood of importing low quality models (global lDDT < 0.5) when compared with the previous QMEAN4 Z-score based criteria.
Membrane Prediction · February 2021
We identify transmembrane proteins in the weekly integration of PDB structures in the SWISS-MODEL Template Library and the SWISS-MODEL Repository.
An implicit solvation model implemented in OpenStructure (mol.alg.FindMembrane) estimates the optimal membrane location for each biounit and performs a classification based on energetic and geometric criteria.
The original algorithm and the used energy function are described by Lomize et al.
The membrane annotations are transfered to models in SWISS-MODEL workspace and SWISS-MODEL Repository.
Structure Comparison · November 2020
The results of homology modelling, structure assessment, QMEAN analysis and structures found in SWISS-MODEL Repository can be combined in this new tool.
The consistency of the selected structures is displayed as a chart, variance heatmap and a new colour scheme on the alignment and 3D viewer. Start a comparison now or visit the help page to learn more.
User Annotations in SWISS-MODEL Repository · November 2020
In SWISS-MODEL Repository, we display UniProt sequence features and InterPro domains from categories which are deemed helpful for structural biologists. These are displayed in the alignment and 3D viewer.
You can now upload your own custom annotations at the residue level. Any sequence-level annotation you can think about can now be displayed in 3D. You can for instance visualize mutation site clusters in 3D.
HHblits3 · September 2020
HH-suite is a widely used open source software suite for sensitive sequence similarity searches and protein fold recognition. It is based on pairwise alignment of profile Hidden Markov models (HMMs), which represent multiple sequence alignments of homologous proteins.
We updated the template search from HHblits 2 at this time with the result being template search results now show more remote homologues.
HHblits3 is enabled for all new projects template search and auto model projects.
GMQE (Global Model Quality Estimation) · September 2020
GMQE (Global Model Quality Estimate) aims to predict the expected quality of a protein model when using a certain template for modelling. It combines properties from the target–template alignment and the template structure.
We introduced a multi-layer perceptron approach to combine these properties and increase prediction accuracy.
SARS-Cov-2 · February 2020
SARS-CoV-2, Severe acute respiratory syndrome coronavirus 2, is a positive-sense, single-stranded RNA coronavirus. It is a contagious virus that causes coronavirus disease 2019 (COVID-19).
We modelled the full SARS-CoV-2 proteome based on the NCBI reference sequence NC_045512 and annotations from UniProt.
The results are available here.
SARS-CoV-2 Variants of Concern, as described by the World Health Organization, were added as Annotation Projects to SWISS-MODEL Repository.
Omicron 21K (BA.1); 21L (BA.2); 22A (BA.4); 22B (BA.5); 22C (BA.2.12.1)
Compare these variants on the spike protein, SPIKE_SARS2
Notable Updates
SWISS-MODEL has been running for many years, and in October 2018 we celebrated 25 years of SWISS-MODEL.
Major recent updates include a new modelling engine (ProMod3, used since June 2016) and the ability to model heteromers where we also introduced major improvements for modelling of homomers (since November 2017).
Also there have been continuous improvements to our template ranking and the local (per-residue) model quality scores.