With the continued global warming, carbon capture has turned into a highly relevant subject to our daily lives. To discover optimal materials for this purpose, we have turned to an innovative method integrating artificial intelligence (AI) and molecular simulations. Leveraging AI and machine learning, over 120,000 new candidates were generated within minutes. High-throughput screening and molecular dynamics simulations were then used to evaluate their stability and carbon capture capacity. As highlighted in a recent publication in Nature Communications, this innovative approach holds potential not only for advancing carbon capture technologies but also for addressing broader challenges in biomolecular simulations and drug design.

Release 3.0 of NAMD retains the same well-known parallel scaling capabilities of previous versions while adding improved support for GPU acceleration. A new GPU-resident mode has been introduced that more than doubles simulation performance for a single GPU and scales more efficiently across NVIDIA DGX architectures than NAMD's GPU-offload mode provided in the earlier version 2.x releases. With GPU-resident mode, NAMD is capable of simulating small systems, such as the 24k-atom DHFR using AMBER force field parameters, at a rate of more than 1 microsecond of simulated time per day. Several advanced features and enhanced sampling methodologies are available for GPU-resident mode, such as alchemical free energy methods, the Colvars collective variables module, and TCL forces.

Permeation of metabolic substrates across biological membranes is a fundamental process in cellular life. This process is largely driven by the concentration gradient of various molecules between the outside and inside of a cell. To meet the need for creating such concentration gradients in MD simulation, and to calculate permeation under natural conditions, we developed a technique in NAMD to continually drive permeant molecules near the periphery of the simulation box across the periodic boundary, which results in a sustained gradient in the center of the simulation system where the membrane is located. This allows for purely diffusive motion of particles across a membrane, enabling one to directly calculate permeability the same way as in experiment. Read more in a recent paper.

The hyperactivity of RAS proteins is associated with tumor progression, invasion, and metastasis in many forms of cancer. RAS proteins must directly interact with the membrane to activate their signaling targets, but the complexity and dynamics of such interactions continue to defy experimental characterization. In a multi-institute collaboration, combining NMR with multi-μs MD simulations with NAMD, and neutron reflectometry, we developed for the first time how a membrane-binding protein domain interacts with and penetrates the surface of the cell at full atomic detail. Analysis of the simulations in VMD revealed that the protein adopts multiple forms on the membrane in a lipid-dependent manner. More details can be found in a recent publication in Nature Communications.

ABCG2 is a membrane transporter regulating the absorption and distribution of over 200 chemical toxins and drugs in the human body. Being able to recognize and transport a wide range of molecules, including a diverse array of chemotherapeutic agents, ABCG2 is one of the main contributors to multidrug resistance in cancer cells. In collaboration with Schuetz lab at St. Jude Children's Research Hospital, and using molecular simulations with NAMD and analyzed by VMD, we showed how a single-point mutation in ABCG2 found in tumor cells cannot complete its transport activity. Our simulations show that formation of a salt-bridge at a critical region due to the mutation may lock the transporter in one structure, thereby preventing it from undergoing conformational changes that are needed for transport. Read more in Drug Resistance Updates.

The latest NAMD 3.0 releases provide GPU-resident molecular dynamics simulation support for external forces, now including Colvars and Tcl Forces. This support allows users to take advantage of a great variety of additional forces in their GPU-accelerated simulations and free energy calculations. The Colvars (collective variables) module and Tcl Forces scripting both provide mechanisms to define external forces between groups of atoms, allowing control over specific structural features during a simulation to enable the study of complex biomolecular processes and interactions. These valuable capabilities are now available from within NAMD's fastest simulation mode.

BmrCD, a multidrug transporter, plays a critical role in drug efflux in bacteria closely related to Staphylococcus aureus. The transporter harvests the energy of ATP to pump drugs out of the cell, thus creating resistance against drugs such as antibiotics. The mechanism of this pumping effect strongly relies on interactions with the lipids in the membrane. To elucidate these underlying protein-lipid interactions, we used VMD to model partially resolved cryo-EM lipids in BmrCD structures solved by the Mchaourab lab at Vanderbilt and simulated their behavior in a bulk membrane using GPU-accelerated NAMD 3.0 at the Center. Simulations revealed that BmrCD engages in an extensive network of interactions with lipids in multiple conformations, elucidating the stabilization of the solved structure. For more details, see our recent publication in Nature Communications.