FUNCLAN - FUNctional annotations through Conformational Landscape Analysis

从与其他蛋白质、小分子（配体）或核酸形成大分子复合物到酶活性在活性和非活性形式之间转换，蛋白质的动态特性导致多种构象状态在许多生物过程中至关重要。为了更好地了解蛋白质的功能，对它们的三维（3D）结构和构象状态进行结构表征至关重要。了解不同的能量有利构象状态之间的转变是理解蛋白质结构和进化原理的基础，有助于解释遗传变异的影响，设计新的药物分子，并在分子水平上阐明耐药性。尽管PDB已经存档了超过165,000个单独的结构，但基于UniProt接入交叉参考的独特蛋白质数量增长速度较慢，总计只有约50,000个，不同序列之间的冗余率存在相当大的差异。这是因为每种蛋白质在PDB中可能有多个代表：配体结合形式和非结合形式；多空间组或样品条件下的结构；与其它大分子（蛋白质或核酸）或由较小的结构域或序列变异决定的结构复杂的。因此，PDB中的结构为理解配体结合位点、单个蛋白质分子以及大型大分子机器的构象灵活性提供了宝贵的资源。利用可用结构集合了解配体结合位点、单个蛋白质分子和大型大分子复合物的异同，有助于破译大分子功能的分子水平细节。不同构象状态数据的可用性也将有助于在全细胞断层摄影中描述颗粒的特征，从而允许在不同疾病或发育状态下的全细胞的分子表型。在这个项目中，我们将增强GESAMT，结构比较算法，以获得配体结合位点，单个蛋白质或结构域和大分子组装的构象灵活性。新的框架FUNCLAN将包括实现有意义的聚类的必要指标，以及描述不同聚类成员之间结构相似性和差异性的必要方案。每个集群都有一个代表性的结构，使用来自pbe - kb的结构和功能注释，我们将描述每个集群并提供生物学背景。新的功能将对已知的数据集进行验证，这些数据集来自大分子和具有特定构象状态的复合物的文献。将实现一个PDB档案范围内的配体结合位点、单个大分子和大分子复合物的聚类管道。结果数据将通过REST API、FTP站点以及一个新的基于web的应用程序以编程方式提供。

The dynamic nature of proteins leading to multiple conformational states is critical in many biological processes from forming macromolecular complexes with other proteins, small molecules (ligands) or nucleic acids to switching between active and inactive forms for enzymatic activity. To gain improved mechanistic insights into the function of proteins, structural characterisation of their three-dimensional (3D) structures and their conformational states is critical. Knowledge of the transition between different energetically favoured conformational states is fundamental to the understanding of the principles of protein structure and evolution and can help in explaining the effects of genetic variants, in designing new drug molecules and in elucidating drug resistance at the molecular level.Although the PDB has archived more than 165,000 individual structures, the number of unique proteins based on the number of UniProt accession cross-references grows at a slower pace and totals only ~50,000, with a considerable variation in the redundancy rate amongst different sequences. This is because each protein may have multiple representatives in the PDB: ligand-bound and unbound forms; structures in multiple space groups or sample conditions; in complex with other macromolecules (proteins or nucleic acid) or structures determined of smaller domains or sequence variants. Thus, the structures in the PDB provide a valuable resource for understanding the conformational flexibility of ligand binding sites, individual protein molecules as well as large macromolecular machines. Understanding the similarities and differences in ligand binding sites, individual protein molecules and the large macromolecular complexes using the ensemble of available structures can assist in deciphering the molecular level details of macromolecular function. The availability of data on distinct conformational states will also assist in characterising the particles in whole-cell tomograms, thus allowing molecular phenotyping of whole cells in different disease or development states.In this project we will enhance GESAMT, the structure comparison algorithm, to derive conformational flexibility of ligand binding sites, individual proteins or domains and macromolecular assemblies. The new framework, FUNCLAN, will include the necessary metrics to realise meaningful clustering and the necessary scheme to describe the structural similarities and differences between members of different clusters. Each cluster will have a representative structure and using the structural and functional annotations from PDBe-KB, we will characterise each cluster and provide biological context. The new functionality will be validated against a dataset of known examples from the literature of macromolecules and complexes exhibiting specific conformational states. A pipeline for a PDB archive-wide clustering of ligand binding sites, individual macromolecules and macromolecular complexes will be implemented. The resulting data will be made available programmatically via a REST API, an FTP site, and also via a novel web-based application.