Investigating sequence-dependent DNA mechanics and nucleosome stability through coarse-grained simulations

DNA sequence encodes genetic information, but also influences DNA shape, deformability and interactions with proteins. In nucleosomes, these properties affect how readily DNA accommodates the curved path imposed by the histone octamer and therefore contribute to sequence-dependent nucleosome stability. In this thesis, sequence-dependent nucleosome behaviour was investigated using coarse-grained molecular dynamics simulations combining MADna for DNA with CALVADOS 3 for the histone octamer. Thermodynamic Integration calculations in DNA-only and histone-associated systems were used to determine relative nucleosome-binding free energies for controlled A-rich variants, randomized controls, Widom 601 and two chromosome XIII sequences. The thermodynamic results were then interpreted using analyses of DNA geometry, nucleosomal wrapping, DNA–histone proximity and sequencedependent crookedness. The clearest effect was observed for the A-rich series: progressively extending the central adenine tract increased the relative nucleosome-binding preference for 1KX5 over the A-rich variants. This trend correlated strongly with the change in crookedness and was accompanied by greater DNA opening and departures from the compact superhelical geometry. The randomized sequences and Widom 601 produced smaller differences that did not support a robust thermodynamic ordering given the numerical and sampling uncertainty indicated by the consistency test. For the chromosome XIII pair, the preference for +920 over −397 agreed qualitatively with the experimental occupancy ordering and was only weakly affected by linker DNA. Overall, the results show that the coarse-grained framework captures the response to strong, controlled sequence perturbations and connects the calculated relative nucleosome-binding free energies with changes in DNA geometry and nucleosomal wrapping. Intrinsic limitations of the coarse-grained models must also be considered when interpreting the molecular origin of subtler sequence-dependent effects. The approach therefore provides a computationally efficient framework for investigating how DNA sequence and mechanics contribute to nucleosome stability, while identifying the sampling and numerical limitations that constrain subtler sequence-dependent effects.