A new class of enhanced kinetic sampling methods for building Markov state models

Area of Science:

• Computational chemistry
• Biophysics
• Materials science

Background:

• Markov state models (MSMs) are crucial for studying long-timescale dynamics in biomolecular and materials systems.
• MSMs are often built using molecular dynamics (MD) simulations, but typically capture only parts of the configurational space, leading to missing states and pathways.
• The finite validity time of MSMs is often shorter than the MD simulation time invested, limiting their predictive power.

Purpose of the Study:

• To present a general framework relating kinetic uncertainty to MSM validity time, missing states/pathways, network topology, and sampling.
• To introduce novel enhanced kinetic sampling techniques to effectively boost MSM validity time.
• To demonstrate the application of these methods using various examples.

Main Methods:

• Development of a theoretical framework connecting kinetic uncertainty, model limitations, and sampling strategies.
• Introduction of enhanced kinetic sampling techniques focused on rare states and pathways.
• Application and validation using 1D energy landscapes, lattice models, and biomolecular systems.

Main Results:

• A framework is established to quantify the factors limiting MSM validity time.
• The study shows that sampling frequently observed states may not improve MSM validity time.
• New enhanced sampling techniques are introduced that effectively target uncertainty-contributing rare states/pathways.

Conclusions:

• MSM validity time is inherently limited by incomplete configurational space coverage and sampling.
• Targeting rare states/pathways that dominate kinetic uncertainty is key to enhancing MSM validity time.
• The presented enhanced kinetic sampling methods offer an effective strategy for improving the accuracy and timescale of kinetic models.