This is the authors' abstract. We don't add key points for this paper.
Path collective variables reduce a high-dimensional transition to a progress coordinate, s, and an orthogonal distance, z. Computing the free energy along s often requires restraining z, so that sampling embraces a tube centered on the reference path. However, a conventional fixed half-harmonic wall demands an a priori wall position that is both system dependent and s-dependent. Too tight a tube biases the path-projected free energy. Conversely, too loose a tube induces numerical instability and inter-channel leakage. We present an adaptive tube restraint, the wall position of which evolves on the fly to track a fixed contour ΔF∗ of the orthogonal free energy F(z∣s), widening automatically in flat basins and narrowing at pinched saddles, using a user-selected transverse energy scale. We prove that, for a locally harmonic perpendicular well, a contour-following tube captures an s-independent fraction of the orthogonal partition function, so that in the hard-wall limit, its bias cancels out. In stark contrast, the bias of a fixed-width tube varies with s and distorts the free-energy landscape. We probe the algorithm on several model potentials, on N-acetyl-N′-methylalanylamide isomerization, on the folding of the mini-protein chignolin, and on the folding-upon-binding of the ribose-binding protein. The method is implemented in the open-source Colvars library and is, therefore, usable within popular MD engines, such as NAMD, LAMMPS, GROMACS, and Tinker-HP.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Talmazan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: