Automatic discovery and annotation of the impact of chemical modifications in RNA on structure and binding

存在于每个生物体中的生物分子是核糖核酸（RNA）。像脱氧核糖核酸（DNA）一样，它们由四种不同的成分组成，称为碱基。它们共享前三种称为腺嘌呤(A)，胞嘧啶(C)和鸟嘌呤(G)。第四种是RNA中特有的，称为尿嘧啶(U)。它们扮演着许多重要的角色，如传递信息、制造蛋白质和调节基因。为了做到这些，他们采用了复杂的三维结构。在过去的几十年里，测序技术为我们提供了大量关于RNA序列集合的信息。许多复杂的实验技术也允许以极高的精度确定一些三维结构。尽管如此，获得结构比序列更具挑战性。计算方法已经被用来理解序列、结构和最终功能之间的关系。事实证明，这些工具对于理解已经产生的大量序列数据非常宝贵，并且可以通过序列中的错误来检测疾病，甚至可以通过创建采用具有理想功能的特定结构的序列来设计新药。我们现在知道，这个序列还不足以完全理解RNA的功能。在这两种情况下，序列之外的表观遗传因素表现为碱基上的化学修饰。最近发现，它们对RNA的功能至关重要，因为它们可以对RNA的结构产生强烈的影响。虽然在晶体结构中观察到超过190种修饰，但只有少数被研究过。自今年以来，他们的数量一直在稳步增长，第一种大规模检测序列化学修饰的测序技术被设计出来，并有望获得大量新数据。本提案旨在创建新的计算方法来了解化学修饰对其结构和功能的影响。它分为两个主题。第一个将研究含有化学修饰rna的已知结构。我们将设计一个新的本体来描述涉及这些化学修饰碱基对的新的相互作用。我们将同时创建新的算法来识别保守的结构模块，特别是那些相互作用的模块，以了解RNA的整体构象是如何受到影响的。这些将用于更好地预测全局构象，并设计具有新颖理想功能的rna。第二个主题将着眼于序列。虽然它们通常没有已知的结构，但它们包含了大量可以从它们的大量中提取的信息。我们将创建生物信息学方法，从共同进化的序列中连接的位置网络中提取进化信息，暗示它们的功能必要性。这些将用于理解和预测化学修饰与其他分子结合的作用。

A bio-molecule present in every living organism is Ribonucleic Acid (RNA). Like Desoxyribonucleic acid (DNA) they are formed of four different components, called bases. They share the first three called Adenine (A), Cytosine (C) and Guanine (G). The fourth is unique in RNA and called Uracil (U). They fulfill many essential roles, as transferring information, making proteins, and regulating genes. To do those things, they adopt complex three-dimensional structures. In the past decades, sequencing technologies has given us a lot of information about the ensemble of RNA sequences. Many complex experimental techniques have also allowed to determine some three-dimensional structures with extreme precision. Nonetheless, obtaining the structure is much more challenging than the sequence. Computational methods have been used to understand the relationship between the sequence, the structure, and ultimately their function. Those tools have proven to be invaluable to understand the massive amount of sequence data that has been produced and allowed to detect diseases by errors in the sequence, even to design new drugs by creating sequences that adopt specific structures with a desirable function. We now know that the sequence is not enough to fully understand the function of RNA. In both cases epigenetic factors, exterior to the sequence, manifest themselves as chemical modification on the bases. It has been recently discovered that they are essential for RNA to perform as they can have a strong effect on their structure. While over 190 modifications are observed in crystal structure, only a handful have been studied. Their number is steadily growing since just this year, the first sequencing technique to massively detect chemical modifications in sequences was designed, and promises a flood of new data. This proposal aims to create new computational methods to understand the impact of the chemical modification on their structure and function. It is divided in two main themes. The first one will study known structures that contain chemically modified RNAs. We will design a new ontology to describe the new interactions involving those chemically modified base pairs. We will in parallel create new algorithms to identify conserved structural modules, especially those with interactions, to understand how the global conformation of the RNA is impacted. Those will be used to better predict global conformation, and design RNAs with novel desirable functions. The second theme will look at the sequences. Although they often are without a known structures, they contain a lot of information that can be extracted from their large numbers. We will create bioinformatic methods to extract from evolutionary information connected networks of positions in the sequences that have co-evolved together, hinting at their functional necessity. Those will be used to understand and predict the role of chemical modifications in binding to other molecules.