Bayesian analysis of time series of single RNA under fluctuating force

TL;DR

This paper develops a Bayesian Monte Carlo method to accurately infer intrinsic kinetic parameters of single RNA molecules from noisy force spectroscopy data, accounting for physical fluctuations in experimental conditions.

Contribution

It introduces a coarse-grain physical model and a Bayesian inference approach that outperform traditional methods in analyzing single-molecule force data.

Findings

Bayesian method improves accuracy over histogram fitting

Model captures key physical factors affecting measurements

Method effectively infers kinetic parameters from noisy data

Abstract

Extracting the intrinsic kinetic information of biological molecule from its single-molecule kinetic data is of considerable biophysical interest. In this work, we theoretically investigate the feasibility of inferring single RNA's intrinsic kinetic parameters from the time series obtained by forced folding/unfolding experiment done in the light tweezer, where the molecule is flanked by long double-stranded DNA/RNA handles and tethered between two big beads. We first construct a coarse-grain physical model of the experimental system. The model has captured the major physical factors: the Brownian motion of the bead, the molecular structural transition, and the elasticity of the handles and RNA. Then based on an analytic solution of the model, a Bayesian method using Monte Carlo Markov Chain is proposed to infer the intrinsic kinetic parameters of the RNA from the noisy time series of the distance or force. Because the force fluctuation induced by the Brownian motion of the bead and the structural transition can significantly modulate the transition rates of the RNA, we prove that, this statistic method is more accurate and efficient than the conventional histogram fitting method in inferring the molecule's intrinsic parameters.