Software

NIH Center For Macromolecular Modeling and Visualization

To investigate biophysics problems at the atomic and subatomic level, our group develops computational tools for Molecular Dynamics, both at the classical and the quantum level. Additionally, our group is part of the NIH Center for Macromolecular Modeling and Visualization, which is known worldwide for the development of both NAMD and VMD software. Particularly, we are on the main development team of VMD's QwikMD, a tool that simplifies the use of Molecular Dynamics tools, and the hybrid QM/MM implementation of NAMD. More recently we also started to work on combining artificial intelligence tools to both NAMD and VMD.

NAMD 3

NAMD 3 is the latest version of the award-winning molecular dynamics engine, built for GPU-resident, high-performance simulations of large biomolecular systems. It leverages full GPU acceleration to minimize data transfer bottlenecks and maximize simulation speed. Deeply integrated with VMD 2, NAMD 3 enables real-time visualization and analysis of simulations as they run. It is distributed free with source code and precompiled binaries. Tutorials are available to help users get started with biomolecular modeling using NAMD and VMD.

VMD 2

VMD 2 is the next-generation tool for modeling, visualization, and analysis of biomolecular systems. It supports standard formats like PDB and offers a wide range of rendering styles. Building on the original VMD, it adds a modern interface, GPU acceleration, and powerful tools for analyzing large-scale molecular dynamics (MD) simulations. VMD 2 also integrates seamlessly with NAMD 3, enabling real-time visualization and analysis of simulations running on local or remote systems.

Hybrid QM/MM Suite

NAMD QM/MM interface extends existing NAMD features to the quantum mechanical level, presenting features that are not yet available in any QM/MM implementation. The first is the ability to execute multiple QM regions in parallel, thorough independent executions of your choice of quantum chemistry code. This allows one to account for multiple reaction centers that are known to work synergistically, for example, or even distant allosteric regulation sites and a reaction center. Investigation of processes occurring on a timescale usually not accessible by QM/MM methods can now be performed by a combination of temperature replica exchange molecular dynamics and QM/MM molecular dynamics. Taking advantage of its integration to VMD, NAMD QM/MM is an easy-to-use platform for hybrid simulations.

Network Analysis

The Generalized Network Analysis tool allows users to study allostery and signaling by mapping dynamic interaction networks onto molecular structures. It supports residue-level or atom-level networks derived from simulations or experiments and enables visualization of communication pathways, identification of key hubs, and comparison across different functional states (e.g., wild-type vs. mutant). Especially useful for analyzing long-range effects and mechanical signaling, the tool has been applied to kinases, adhesins, and other mechanoactive proteins.

QwikMD

QwikMD is a VMD plugin to help to start and analyze MD simulations. The plugin helps, specially scientists that are starting to perform MD simulations, to prepare the necessary files to run these simulations in desktop machines all the way to large supercomputers. All the necessary steps, from the PDB to the configuration file is created with simple procedures so the user one can use the plugin to learn how to prepare MD simulations. The live simulation option allows for the visualization and analysis of the simulation on the fly, helping new users to learn more about MD simulations and expert users to test their simulations before submitting it to run in a supercomputer.

QwikFold

QwikFold is graphical interface plugin for artificial intelligence (AI) based structural biology tools, such as AlphaFold.

Artificial Intelligence

Analyzing large sets of Molecular Dynamics (MD) trajectories is a major problem in computational biophysics. Combining extensive MD simulations to create a unique database allow us to link protein dynamics to mechanostability and binding affinity. We are using this database to inform the development and optimization of artificial intelligence (AI) models for optimization of protein interfaces.