Source-linked AI summary
CABS-flex 2.0: a web server for fast simulations of flexibility of protein structures
Aleksander Kuriata, Aleksandra Maria Gierut, Tymoteusz Oleniecki, Maciej Pawel Ciemny, Andrzej Kolinski, Mateusz Kurcinski, Sebastian Kmiecik
TL;DR
Protein-flexibility simulations are computationally demanding, particularly for large proteins, while existing methods and systems have design- and scope-dependent limitations. CABS-flex 2.0 updates the fast coarse-grained web server with larger and multimeric protein support, customizable restraints and simulation parameters, contact maps, and an enhanced interface. The update broadens the server's supported simulation workflows while retaining principles validated against atomistic MD and NMR data.
Problem
Protein-flexibility simulations remain computationally demanding for large proteins, and computational methods have limitations depending on their design and modeled system.
Method
CABS-flex 2.0 uses a coarse-grained CABS simulation pipeline with customizable simulation settings and distance restraints, followed by representative-model reconstruction to all-atom representation.
Results
The updated server supports proteins up to 2,000 amino acids and up to 10 chains, and adds contact maps plus enhanced analysis features.
Takeaways & Limitations
CABS-flex 2.0 broadens fast protein-dynamics simulations to larger, multichain, and more specialized systems through user-tailored controls and analysis outputs.
Takeaways & Limitations
Default settings and restraints were optimized for convergence with fluctuations from atomistic MD simulations of globular proteins and were correlated with NMR ensembles.
Abstract
from arXiv · showhide
Classical simulations of protein flexibility remain computationally expensive, especially for large proteins. A few years ago, we developed a fast method for predicting protein structure fluctuations that uses a single protein model as the input. The method has been made available as the CABS-flex web server and applied in numerous studies of protein structure-function relationships. Here, we present a major update of the CABS-flex web server to version 2.0. The new features include: extension of the method to significantly larger and multimeric proteins, customizable distance restraints and simulation parameters, contact maps and a new, enhanced web server interface. CABS-flex 2.0 is freely available at http://biocomp.chem.uw.edu.pl/CABSflex2
INTRODUCTION
CABS-flex 2.0 substantially expands fast protein-flexibility simulations to larger and multichain proteins while adding customizable controls, contact maps, and an enhanced interface.
- Classical atomistic simulations of biologically relevant protein fluctuations are computationally demanding, motivating coarse-grained simulation with all-atom reconstruction.
- CABS-flex 2.0 extends supported inputs from 400 to 2,000 amino acids and from single-chain proteins to proteins of up to 10 chains.
- Contact maps report residue-residue contact frequencies during simulation and for 10 representative protein models, supporting analysis of modeled interactions.
- The updated server addresses original size limitations through larger computational resources and rewritten code, while retaining principles validated against atomistic MD and NMR data.
- Customizable temperature, simulation length, distance restraints, and related options enable tailored simulations of disordered regions, flexible loops, and user-defined restraints.
- The pipeline uses the CABS coarse-grained model, clusters a dynamics trajectory into 10 representative models, and reconstructs them automatically to all-atom representation.
DESCRIPTION OF THE WEB SERVER
CABS-flex 2.0 accepts protein structures with multiple chains and offers extensive control over restraints, flexibility assignments, and simulation settings. Its web interface supports customized simulations through automatic and manually edited restraints.
- Protein structure input: The server accepts PDB codes or uploaded PDB files, including multiple chains, with complete backbone atoms required for every residue.Each chain may contain up to 2000 residues.
- Simulation controls: Users can customize temperature, simulation length, distance restraints, restraint modes, residue gaps, and minimum and maximum restraint distances.Default minimum and maximum distances are 3.8 and 8.0 Å, respectively.
- Simulation controls: Residue flexibility values range from 0 for fully flexible to 1 for rigid and weight restraints using the lower flexibility of each restrained residue.Flexibility can be assigned uniformly, from beta factors, or through custom residue or range rules.
- Restraint editing: Generated restraints can be manually edited, deleted, or supplemented with new Cα–Cα and side-chain–side-chain restraints.Additional restraints may be entered in a text box or uploaded from a text file.
- Simulation controls: Advanced options control trajectory generation: the saved trajectory contains 20 x Ncycle models, while Nskipped determines how many generated models separate saved frames.For Ncycle = 50 and Nskipped = 100, 1000 of 100000 generated models are written to the trajectory.
Output files and data
Completed jobs provide interactive model visualization, contact maps, and residue-wise fluctuation plots. Results can be inspected through selectable structure sets and downloaded as structural, graphical, or numerical data.
- Output tabs: Completed jobs add Models, Contact maps, and Fluctuation plot tabs.These tabs organize structural models, interaction patterns, and residue-level dynamics.
- Models: The Models tab displays 10 final structures and supports interactive viewing, selection, downloading, surface or cartoon representations, and rotation.The model_all set visualizes all 10 models together to show final-model heterogeneity.
- Contact maps: Contact maps encode residue-residue interactions, with colors representing the frequency of each contact across the selected structure set.Users can inspect individual final models or the trajectory and adjust the displayed contact range.
- Fluctuation plot: The Fluctuation plot reports residue-wise RMSF after global superposition and provides downloadable SVG graphics and CSV data.Multichain proteins receive a separate plot for each chain.
SERVER ARCHITECTURE AND DOCUMENTATION
CABS-flex 2.0 uses a Flask and Jinja2-generated HTML interface backed by MySQL storage and queued server jobs. Online documentation is provided through the server’s “How to” page.
- Server architecture: The server dynamically generates its HTML interface with Flask and the Jinja2 templating engine.Validated user data are stored in a MySQL database.
- Server architecture: Jobs start when resources are available or enter a queue, with progress reported through pending, in queue, running, and done statuses.The interface reports job status to users during computation.
- Documentation: Online documentation for CABS-flex 2.0 is available under the “How to” subpage.
SUMMARY
CABS-flex 2.0 is a web server for efficient coarse-grained simulations of large-scale protein and protein-complex fluctuations. It expands system size and flexibility control while producing mobility profiles and trajectories for downstream biomolecular analyses.
- Summary: CABS-flex 2.0 enables efficient simulations of large-scale structure fluctuations in proteins and protein complexes.The method builds on coarse-grained simulations used in the earlier CABS-flex server and related modeling tools.
- Summary: The updated server supports larger systems and more elaborate control of amino-acid-chain flexibility than the previous implementation.Users can designate restricted regions and regions that may adopt structures differing substantially from the input.
- Applications: Mobility profiles and fluctuation trajectories can support identification of mobile fragments, receptor-structure generation for docking, and studies of allosteric equilibria.The paper also mentions other tasks requiring large-scale dynamics data.
- Availability: The server is freely available online at http://biocomp.chem.uw.edu.pl/CABSflex2.