Constructing multi-resolution Markov State Models (MSMs) to elucidate RNA hairpin folding mechanisms

Area of Science:

• Computational Chemistry
• Biophysics
• Molecular Dynamics

Background:

• Simulating biological processes at atomic resolution requires capturing long timescales, which is computationally challenging for standard atomistic simulations.
• Markov State Models (MSMs) offer a solution by extracting long-time dynamics from shorter simulations, effectively coarse-graining conformational space into metastable states.
• MSMs are inherently multi-resolution, allowing for varying degrees of detail in analyzing molecular dynamics.

Purpose of the Study:

• To introduce a novel algorithm, Super-level-set Hierarchical Clustering (SHC), for constructing multi-resolution Markov State Models.
• To demonstrate the capability of SHC in generating MSMs at different resolutions.
• To analyze the folding dynamics of a small RNA hairpin using the developed multi-resolution MSMs.

Main Methods:

• Development of the Super-level-set Hierarchical Clustering (SHC) algorithm, which clusters super levels of phase space density.
• Application of SHC to atomistic simulation data of a small RNA hairpin.
• Generation and validation of Markov State Models at four distinct resolutions.

Main Results:

• The SHC algorithm successfully generated multi-resolution MSMs that accurately reproduce the original simulation data.
• Analysis of the RNA hairpin folding dynamics revealed no metastable on-pathway intermediate states.
• The folded state was identified as a central hub directly connected to various unfolded/misfolded states.

Conclusions:

• The SHC algorithm provides an effective method for constructing multi-resolution Markov State Models.
• The folding pathway of the small RNA hairpin is characterized by direct transitions from the folded state to multiple unfolded states, bypassing stable intermediates.
• This study advances the capability to simulate and understand complex molecular dynamics over biologically relevant timescales.