Effect of base-pair sequence on B-DNA thermal conductivity

TL;DR

This study uses molecular dynamics simulations to explore how different DNA base-pair sequences affect thermal conductivity, revealing insights relevant for DNA-based thermal device engineering.

Contribution

It systematically investigates the influence of various DNA sequences and lengths on thermal conductivity, including mixed sequences, using RNEMD simulations, which is novel in this context.

Findings

No significant divergence in thermal conductivity at lengths up to 80 base pairs.

Short encapsulated AT sequences within CG sequences increase thermal conductivity.

Thermal conductivity follows a power law with length, but does not violate Fourier law at studied lengths.

Abstract

The thermal conductivity of double-stranded (ds) B-DNA was systematically investigated using classical molecular dynamics (MD) simulations. The effect of changing base-pairs on the thermal conductivity of dsDNA, needed investigation at a molecular level. Hence, four sequences, viz. poly(A), poly(G), poly(CG) and poly(AT) were initially analysed in this work. Firstly, length of these sequences was varied from 4-40 base-pairs (bp) at 300 K and the respective thermal conductivity ($\mathrm{\kappa}$) was computed. Secondly, the temperature dependent thermal conductivities between 100 K and 400 K were obtained in 50 K steps at 28 bp length. The M\"{u}ller-Plathe reverse non-equilibrium molecular dynamics (RNEMD) was employed to set a thermal gradient and obtain all thermal conductivities in this work. Moreover, mixed sequences using AT and CG sequencces, namely $\mathrm{A(CG)_{n}T}$ (n=3-7), $\mathrm{ACGC(AT)_{m}GCGT}$ (m=0-5) and $\mathrm{ACGC(AT)_{n}AGCGT}$ (n=1-4) were investigated based on the hypothesis that these sequences could be better thermoelectrics. 1-dimensional lattices are said to have diverging thermal conductivities at longer lengths, which violate Fourier law. These follow power law, where $\mathrm{\kappa \propto\ L^{\beta}}$. At longer lengths, the exponent $\mathrm{\beta}$ need to satisfy the condition $\mathrm{\beta}>1/3$ for divergent thermal conductivity. We find no such significant Fourier law violation through divergence of thermal conductivities at 80 bp lengths or 40 bp lengths. Also, in the case of second study, the presence of short (m$\le$2) encapsulated AT sequences within CG sequences show an increasing trend. These significant results are important for engineering DNA based thermal devices.