Bridging Disparate Structural/Functional Scales: Multiscale Modeling of the Chromatin Fiber and RNA Tertiary Structures

摘要 今天，实验生物学和计算生物学的进步正在帮助生物学和生物医学的许多领域取得令人兴奋的发现。物理实验和建模共同解决了生物学中的基本难题，并通过分子设计、靶向治疗和纳米技术将这些发现应用于医学和技术。PI的计算生物学/生物物理实验室专注于与实验者合作，利用数学、计算机科学、工程以及生物和化学的想法，通过创新的分子模型和计算方法，解决关于核酸及其复合体(特别是染色质和RNA)的基本结构和动力学问题。这个MIRA项目将整合NIGMS关于染色质结构、RNA设计和RNA结构预测的三个项目，以促进我们对染色质和RNA结构/功能关系的基本理解，使用多尺度模型连接不同的尺度，以允许在染色质和RNA折叠方面进行前所未有的应用。对于染色质，从全基因组实验中获得的关于核小体和远程接触图的详细原子信息将被一系列可调的粗粒度模型连接起来，以连接原子、中尺度和聚合物模型，以研究癌细胞中基因环机制对染色质高阶结构的表观遗传调节；这些用于抑制/激活转录的结构机制具有翻译分支，通过通过染色质环溶解来有针对性地重新表达那些沉默的基因，用于癌症治疗。对于RNA，通过利用2D图的急剧变量减少和自然模块化来解决RNA和RNA/蛋白质三级基序预测不佳的一般问题，将有效的粗粒度图采样、图论子结构和数据挖掘与我们的程序ragtop中对伪结、扭结转折和RNA-蛋白质结合基序的改进处理相结合。我们将为社区的努力做出贡献，如RNA拼图；为预测的但尚未发现的类似RNA的折叠基序设计新的RNA，以产生用于设计的模块图集；并设计具有所需荧光特性的核糖开关，如传感器。这些工作在染色质和RNA结构和功能的领先专家的支持下，将有助于推进生物/生物医学的基本领域，包括基因组生物物理学、染色质高级结构、基因表达以及细胞发育和分化，从而有针对性地治疗与基因表达异常相关的人类疾病，包括癌症、遗传疾病和退行性疾病。由此产生的多尺度计算范例广泛适用于其他生物分子过程，并将与广大社区共享。

Abstract Advances in both experimental and computational biology are helping make exciting discoveries in many fields of biology and biomedicine today. Together, physical experiments and modeling are solving fundamental puzzles in biology as well as applying the findings to medicine and technology through molecular design, targeted therapies, and nanotechnology. The PI's computational biology /biophysics lab focuses on solving fundamental structural and dynamical questions concerning nucleic acids and their complexes (notably chromatin and RNA) in collaboration with experimentalists by innovative molecular models and computational methods using ideas from mathematics, computer science, and engineering, as well as biology and chemistry. This MIRA project would consolidate three NIGMS projects on chromatin structure, RNA design, and RNA structure prediction to advance our fundamental understanding of structure/function relationships for chromatin and RNA using multiscale models that bridge disparate scales to allow unprecedented applications regarding chromatin and RNA folding. For chromatin, the detailed atomic information on nucleosomes and long-range contact maps available from genome-wide experiments will be bridged by a hierarchy of tunable coarse-grained models to link atomic, mesoscale, and polymer models to investigate the epigenetic modulation of chromatin higher-order structure by gene looping mechanisms in cancer cells; these structural mechanisms for repression/ activation of transcription have translational ramifications through targeted re-expression of those silenced genes by chromatin loop dissolution for cancer therapy. For RNA, the general problem of poor predictions of RNA and RNA/protein tertiary motifs will be tackled by exploiting the drastic variable reduction and natural modularity of 2D graphs to combine efficient coarse-grained graph sampling, graph theory substructuring, and data-mining with improved handling of pseudoknots, kink-turns, and RNA-protein-binding motifs in our program RAGTOP. We will contribute to community efforts like RNA-puzzles; design novel RNAs for predicted, but yet undiscovered, RNA-like fold motifs to produce a atlas of modules for design; and design riboswitches with desired fluorescent properties, as sensors. These works, with support from leading experts in chromatin and RNA structure and function, will help advance fundamental areas of biology /biomedicine including genome biophysics, chromatin higher-order structure, gene expression, and cell development and differentiation and hence the targeted treatment of human diseases associated with aberrant gene expression, including cancers, genetic disorders, and degenerative diseases. The resulting multiscale computing paradigms are widely applicable to other biomolecular processes and will be shared with the community at large.