Software
designed for my research
designed for my research
Welcome to the software section of my website, where I present the various scientific research software packages that we have developed in my group and make available to the community. My goal is to provide original and user-friendly tools that can help researchers in their work. Whether you are a seasoned professional or a student just starting out, I sincerely hope you will find something of value here. Browse through this selection and feel free to contact me with any questions or suggestions. Thank you for visiting!
Here is a matrix with shortcuts to overviews for each of the packages:
After a brief description of each package, you will find important links to builds, source code and websites, as well as licensing information. Relevant literature references are also listed. If you use any of our code, please let us know. If you improve or change it, please share your developments with us and help us make the software better.
UnityMol is a molecular viewer and prototyping platform developed by a dedicated team in my lab. It is coded in C# using the Unity3D game engine and was completely redesigned in 2019 to provide a clean code base and new features. The main focus is on the VR version, which supports HTC Vive and Oculus headsets. Other headsets should be supported but have not been tested yet.
The software can read various file formats, including PDB, mmCIF, GRO, Mol2, XYZ, SDF, OpenDX potential maps and XTC trajectory files. It includes HyperBalls, a feature that uses GPU graphics cards and shaders (GLSL or Cg) to visualize molecular structures. The newer versions of UnityMol use a SMCRA (Structure/Model/Chain/Residue/Atom) data structure that supports loading multiple molecules and has a built-in Python console, a selection language adapted to MDAnalysis, various molecule representations, and a modern UI.
The project is actively developed and maintained with the goal of constantly improving its usability and features. As a result, UnityMol is open-source software that is compatible with multiple platforms, including Windows, macOS, Linux, and web apps, and can generate publication-quality images.
In addition, a FAIR UnityMol prototype has been developed to enable sharing of molecular visualizations via the cloud and collaborative virtual reality exploration in immersive 3D environments. This approach to sharing molecular visualizations is based on the principles of FAIR and makes structural and simulation data accessible to a wide audience.
UnityMol is specifically tailored to the different 3D structures of complex carbohydrates and polysaccharides. It was developed using advanced technologies from the video game industry and includes specific functions for the identification and classification of monosaccharides, conformations, arrangements in single or multiple branched chains, the representation of secondary structural elements and essential components in very complex structures.
In summary, UnityMol is a versatile molecular visualization software that supports various file formats, has advanced molecular structure visualization capabilities, and includes a general software architecture for performing interactive molecular simulations in a game engine environment. The software will be actively developed and maintained, with a focus on VR. It is expected to support sharing of molecular visualizations via the cloud in the near future.
| Latest build | 1.1.4 |
| Platforms | Windows, MacOS, Linux |
| Source code Releases | Github |
| License | Dual (GPL + proprietary) |
(under construction).
MDDriver is a software library that facilitates the implementation and use of interactive molecular simulations (IMS), particularly interactive molecular dynamics (IMD) simulations. IMD simulations are a type of interactive simulation technique that enables the study of complex molecular interactions and can be useful in exploring and generating hypotheses about the structural and mechanical aspects of biomolecular interactions. MDDriver makes it easy for any particle-based molecular simulation engine to become interactive.
Interactive molecular simulations are characterized by two main features: the ability to visualize a running simulation in real time and the ability to interactively manipulate the simulation by imposing a force, changing a biophysical property, or editing runtime parameters on the fly. These simulations are not yet widely used in computational biology, but they enable more efficient processing of time-consuming tasks such as modeling complex biomolecular structures or supporting rational drug design.
The MDDriver library is based on NAMD's IMD protocol and uses Gromacs, OPEP, HiRe-RNA, and BioSpring software as example implementations. It allows easy creation of a network between a molecular user interface and a physics-based simulation, providing real-time control and visualization of a running molecular simulation. As an extension, MDDriver was used to develop a VR framework for such immersive and interactive molecular simulations. This framework is based on MDDriver and UnityMol and provides even more interactive capabilities.
One of the most important features of MDDriver is its ability to perform coarse-grained IMD simulations at low resolution. This simplified modeling method is well suited for interactive experiments and provides a good balance between computational speed and modeling accuracy compared to higher resolution all-atom simulations. This is particularly useful for initial exploration and hypothesis development for rare molecular interaction events.
MDDriver has been used to study a variety of biochemical systems, including the enzyme guanylate kinase (GK), the outer membrane protease T, and the soluble N-ethylmaleimide-sensitive factor-attachment protein-receptor complex involved in membrane fusion. The software was used to trigger large conformational changes, perform interactive docking experiments, study lipid-protein interactions, and capture the mechanical properties of a molecular model.
In summary, MDDriver software is a powerful tool that facilitates the implementation and application of interactive molecular dynamics (IMD) simulations. It allows easy creation of a network between a molecular user interface and a physics-based simulation, so that any particle-based molecular simulation engine can become interactive. It is particularly useful for initial exploration and hypothesis development for rare molecular interaction events and has been used to study a variety of biochemical systems. The software has also been used to develop a VR framework for immersive and interactive molecular simulations that provides even more powerful visualization and interaction capabilities.
Epock is an efficient command-line tool for calculating pocket volumes from MD trajectories. It has been specifically optimized for large datasets and is an indispensable tool for drug development, where binding pocket characterization is a key issue often addressed using molecular dynamics simulations (MD).
Epock takes a topology and a MD trajectory as input and uses a configuration file to define the parameters for each cavity to be characterized. The maximum enclosing region (MER) is used to define spatial boundaries, and the approach uses a combination of simple three-dimensional objects (spheres, cylinders, and cuboids) to determine a complex final shape. The software also supports a variety of input formats for trajectories using the VMD plugin.
For each pocket, Epock calculates the space accessible to a probe and provides a detailed analysis of the pocket's volume changes over time. The program was tested with applications for the GLIC ion channel and for characterizing HSP90 binding pockets, and proved to be the fastest program in its category, analyzing the 800-frame GLIC trajectory in just 6 seconds. The analysis provided by Epock highlights the opening of the binding pocket and identifies the specific residues involved in the volume changes of up to 200 Ã….
Overall, Epock is a powerful tool that enables efficient and accurate analysis of large data sets from molecular dynamics simulations, making it a practical tool for researchers and drug developers.
HyperBalls is an efficient software for molecular visualization. It uses the capabilities of graphics processing units (GPUs) to develop new and innovative representations of molecules. The primary representation used in HyperBalls is the "hyperboloid," which seamlessly connects atomic spheres and replaces the traditional cylinders used in other molecular visualization programs. This new representation has many advantages over traditional representations, such as the ability to smoothly represent dynamic bond evolution, which is not possible with cylinders. In addition, HyperBalls can be customized to represent a variety of molecular models, including coarse-grained models of spring networks and ion coordination. This makes it a versatile tool for scientists in a variety of fields.
The core of HyperBalls is implemented using ray-casting GPU shaders. Therefore, one of the biggest advantages of HyperBalls is its performance and pixel-perfect precision. It is currently an efficient way to display a large number of atoms in VR headsets with high image quality. Moreover, its performance increases with the appearance of new generations of graphics cards, without requiring any changes to the source code.
HyperBalls is implemented in UnityMol and designed to be user-friendly, with a simple and intuitive interface. It is also available in several other visualization packages, such as NGL viewer and Sansom.
VItAMInS (Visual and In Situ Analytics for Molecular Interactive Simulation) is a framework that enables the development of in situ and post-mortem applications for the analysis of biological systems. It is based on the component-based middleware FlowVR, which makes it flexible and easy to add new modules or use on different platforms such as personal computers, HPC clusters and virtual reality environments.
VItAMInS enables the creation of a simplified post-mortem analysis application, as we have prototyped with the TrajReader and TrajPlayer modules for the study of molecular dynamics trajectories. Communication with the HyperBalls visualization module is via a FIFO (first in, first out) method. This enables real-time visualization of molecular dynamics simulations and provides an interactive and immersive experience.
In addition to post-mortem analysis, VItAMInS enables the development of in-situ applications. This can be achieved by replacing the TrajReader and TrajPlayer modules with Gromacs and an interface that controls Gromacs. However, this replacement requires several changes in the FlowVR network, since Gromacs is a parallel module that can run on several thousand cores, which means that communication to and from this module must be done via parallel schemes.
The VItAMInS framework was developed to facilitate the creation and execution of simulations and the analysis of the results. It provides a toolset for visualization and analysis of biological systems. The VItAMInS demo is available for download and more details about the pipeline are described in the corresponding papers.
QuickSES is a library for the fast computation of Solvent Excluded Surfaces (SES) on GPUs. This innovative method was developed to provide a fast and open-source solution for computing SES meshes that is accessible to a wide range of users. With QuickSES, you can harness the power of multicore CPUs and massively parallel GPU architectures to render dynamic datasets in interactive time.
Unlike traditional CPU methods that run on a single core, QuickSES uses grid-based techniques that allow easy adaptation to varying performance constraints. This allows you to quickly generate an SES mesh that can then be further processed by mesh-centric tools such as Maya and Unity3D. The QuickSES library can handle large molecular systems.
The QuickSES method is designed for efficient use of memory and can process large molecules quickly. With its advanced CUDA kernels and optimized memory access, QuickSES is able to achieve significant speedup. The library is designed to be very flexible and easy to integrate with existing molecular visualization software. This makes it a good tool for use in drug discovery, rational drug design, and other applications that require fast and accurate calculation of SES.
The BioSpring software enables interactive and flexible docking of large biomolecular assemblies. It is based on an extended elastic network model (aENM) that combines the spring network with unbound terms between atoms or pseudoatoms. This approach allows molecular assemblies to be created on a desktop or laptop computer, thanks to code optimizations including parallel computing and GPU programming. The software greatly simplifies the design of multi-scale scenarios by combining atomistic and coarse-grained models. The software was used to study the filamentous central domain of dystrophin, which spans repeats 11 to 17 and offers insights into medically relevant discoveries.
As part of BioSpring and in collaboration with UnityMol and MDDriver, we have developed an algorithm that performs interactive molecular simulations of protein alignment in membranes and allows monitoring and manipulation of such molecular systems at multiple levels. This is based on the integration of an implicit membrane model for Integral Membrane Protein and Lipid Association (IMPALA) into our IMS framework. Our implementation can cover multiple levels of representation, and the degrees of freedom can be tuned to optimize the experience. Validation of this model in an interactive and exhaustive search mode was explained.
The main goals of BioSpring are to provide a tool for an interactive, rapid overview of biomechanical properties, to assist the user in the complex task of modeling large biomolecular complexes, to perform interactive flexible docking, and to provide and explore new hypotheses for complex molecular relationships. In particular, the positioning of proteins in model membranes can now be performed interactively and in real time using BioSpring.
CC BY-NC-SA 4.0