Computational modeling of sequence-dependent DNA curvature using hydrodynamics

Gonzalez0706951研究人员开发了一种计算方法，从较短、相对较硬的碎片上的流体动力学数据预测DNA的局部、依赖于序列的曲率参数。该方法包括最小化最小二乘泛函，该泛函量化了给定碎片集合的理论和实验沉积速度之间的差异，其数值实现需要围绕各种细长的三维区域重复求解外部斯托克斯型问题。研究工作集中在三个方面：(1)研究具有不同渐近极限的热涨落的Stokes类流体中刚性聚合物的理论沉降速度；(2)基于核自适应Nystrom方法的新思想，设计和分析计算理论沉降速度的快速、可证明收敛的边界元方法；(3)设计和分析在曲率参数依赖于序列的空间上最小化最小二乘沉降函数的快速、可靠的方法。DNA依赖于序列的曲率和灵活性对于它在细胞内的包装、被其他分子识别以及在生物化学重要过程中的构象变化至关重要。然而，很少有实验方法可以用来在基线水平上探索这些性质。该项目的目标是利用DNA在短长度尺度上的刚性，并开发一种从短片段上的流体动力学数据估计局部的、依赖于序列的曲率参数的方法。这种方法是基于观察到的不同形状的相对坚硬的小碎片，当强迫通过粘性流体或稀凝胶时，通常以不同的速度移动。研究人员试图在形状和速度之间建立一种非常精细的数学关系，并进一步利用这种关系开发一种从速度测量中估计形状参数的计算方法。这项研究的结果导致了新的工具来量化DNA的内在序列依赖的曲率，这反过来可能导致更好地理解重要的基因组区域以及DNA如何与其他分子相互作用。

Gonzalez0706951 The investigator develops a computational method forestimating local, sequence-dependent curvature parameters for DNAfrom hydrodynamic data on short, relatively stiff fragments. Themethod consists of minimizing a least-squares functional whichquantifies the difference between theoretical and experimentalsedimentation speeds for a given collection of fragments, and itsnumerical implementation requires the repeated solution of anexterior Stokes-type problem around various slender,three-dimensional domains. The research effort is centered onthree areas: (1) the study of the theoretical sedimentation speedof a stiff polymer in a Stokes-type fluid with thermalfluctuations in different asymptotic limits, (2) the design andanalysis of fast, provably convergent boundary element methodsfor computing the theoretical sedimentation speed based on thenovel idea of a kernel-adapted Nystrom method, and (3) the designand analysis of fast, reliable methods for minimizing aleast-squares sedimentation functional over a space ofsequence-dependent curvature parameters. The sequence-dependent curvature and flexibility of DNA iscritical for its packaging into the cell, recognition by othermolecules, and conformational changes during biochemicallyimportant processes. However, few experimental methods areavailable for probing these properties at the basepair level. The objective of this project is to exploit the stiffness of DNAat short length scales and develop a method for estimating local,sequence-dependent curvature parameters from hydrodynamic data onshort fragments. The method is based on the observation thatsmall, relatively stiff fragments of different shape typicallytravel at different speeds when forced through a viscous fluid orthrough a dilute gel. The investigator seeks to develop asufficiently refined mathematical relation between shape andspeed and further use this relation to develop a computationalmethod to estimate shape parameters from measurements of speed. Results from this research lead to new tools for quantifying theintrinsic sequence-dependent curvature of DNA, which in turn maylead to an enhanced understanding of important genomic regionsand of how DNA interacts with other molecules.