Introducing MSTBx v0.5: Integrating Automated Biomolecular Assembly and AI-Guided Proteome Modeling

Molecular dynamics (MD) simulations are fundamental tools in computational
biophysics, enabling detailed exploration of biomolecular behavior through a
trajectory obtained from solving classical Newton’s equations of motion. The advent
of GPU acceleration has expanded MD’s capabilities, allowing the study of large and
complex systems that complement experimental approaches. However, preparing
initial systems with precise conditions remains a time-consuming and technically
challenging step. To overcome this, we developed MSTBx, a Python-based toolkit
that automates and accelerates the assembly of diverse biomolecular systems by
integrating CHARMM-GUI’s strengths with the flexibility of PSFGen. MSTBx
efficiently builds proteins in solution, membrane environments, protein-ligand
complexes, and large viral assemblies, such as a dengue virus capsid of
approximately 17 million atoms constructed in under two hours. The latest extension
includes MDTranslate.py, a module that converts NAMD PSF/PDB files into
GROMACS-compatible topologies and coordinates, enhancing cross-platform
workflows. Additionally, MSTBx incorporates ColabFoldPipeline.py, a wrapper
enabling automated proteome-scale structure prediction using ColabFold. For
improved structural analysis, MSTBx generates visualization-ready files compatible
with ChimeraX and PyMOL, applying AlphaFold-inspired color schemes to facilitate
interpretation. This unified, modular platform streamlines MD system preparation, file
format translation, proteome modeling, and visualization, establishing MSTBx as an
comprehensive and multi-purpose tool capable of supporting biomolecular
simulations from atomic details to cellular-scale assemblies.