Comprehensive Kinetic and Free Energy Analysis of NTL9 Folding via Systematic Collective Variable Selection and Markov State Models

Abstract: Understanding the complex pathways and kinetics of protein folding from molecular dynamics simulations requires sophisticated analytical tools. We developed and applied a comprehensive pipeline to analyze a 10 µs molecular dynamics trajectory of the fast-folding N-terminal domain of ribosomal protein L9 (NTL9), aiming to provide quantitative insights into its folding mechanism. Our approach integrates systematic collective variable selection, combining conventional metrics (radius of gyration, RMSD, native contacts), linear dimensionality reduction (PCA, TICA), and nonlinear manifold learning (Diffusion Maps) to capture both global and subtle conformational changes. Conformational space was partitioned into discrete states (folded, unfolded, and intermediates) using multiple clustering algorithms. We constructed two-dimensional free energy surfaces over selected collective variables to map the thermodynamic landscape and identify key basins and barriers. Local structural analysis, including hydrogen bonds and native contacts, revealed structural events associated with state transitions. Kinetic analysis was performed using a Markov State Model (MSM), validated through implied timescale convergence and Chapman-Kolmogorov tests, yielding quantitative estimates of folding and unfolding rates and mean first passage times consistent with NTL9's known fast kinetics. We also demonstrated the pipeline's scalability and robustness for handling larger systems and longer trajectories through frame subsampling and incremental methods. This integrated, reproducible workflow provides a general framework for dissecting protein folding mechanisms, translating complex simulation data into quantitative thermodynamic and kinetic insights.

Subjects:	q-bio.BM; q-bio.QM
Cite as:	PX:2508.00021