Integrated Cheminformatics Resource for Orphan Neurodegenerative Diseases

描述（由申请人提供）：在本项目中，我们寻求继续开发新的创新计算工具，将其应用于孤儿神经退行性疾病相关蛋白靶点的建模，并建立一个可公开访问的网络资源来托管这些工具和模型。这些目标将通过两个具体目标来实现。目标1是开发和验证两种新的基于结构的计算工具。第一个工具（Shape4）是一种快速的、基于结构的虚拟筛选方法，用于搜索大型多构象分子数据库，以寻找蛋白质靶标的潜在配体。它是基于这个主要假设：配体分子的地形形状和药效团特征应该与其蛋白质结合位点的特征互补。将采用新的计算几何和形状建模算法来完成形状/药效团匹配任务。第二个工具（SB-PPK）生成基于结构的有机分子描述符。由此产生的描述符不仅取决于有机分子的结构，还取决于目标蛋白的结合位点特征。因此，这些新的描述符克服了传统分子描述符的缺点，即只依赖于有机分子的结构，而不管目标蛋白质是什么。新的描述符将与QSAR（定量结构活性关系）建模工作流程结合使用，以开发选定蛋白质靶点的预测模型（Aim-2a）。具体来说，Aim-1中开发的两种新方法将用于建立来自磷酸二酯酶（PDE）和组蛋白去乙酰化酶（HDAC）家族的预测模型：PDE- 4、PDE-5、HDAC-7和HDAC-8。然后，我们将通过门户网站（Aim-2b）部署经过验证的模型，以使孤儿神经退行性疾病的研究界受益。我们的目标是开发和应用创新的计算工具来研究孤儿神经退行性疾病的蛋白质靶点，并建立一个开放获取的信息学资源来支持这些疾病领域的药物发现工作。我们将最终为减轻孤儿神经退行性疾病患者的痛苦做出贡献。

DESCRIPTION (provided by applicant): In this project, we seek to continue our development of new innovative computational tools, apply them to model relevant protein targets of orphan neurodegenerative diseases, and build a publicly accessible web resource to host these tools and models. These goals will be achieved via two Specific Aims. Aim-1 is to develop and validate two new structure-based computational tools. The first tool (Shape4) is a fast, structure-based virtual screening method designed to search large multi-conformer molecular databases for potential ligands for a protein target. It is based on this chief hypothesis: a ligand molecule's topographical shape and pharmacophore features should be complementary to those of its protein binding site. Novel computational geometry and shape modeling algorithms will be employed to fulfill the shape / pharmacophore matching tasks. The second tool (SB-PPK) generates structure-based descriptors for organic molecules. The descriptors so generated depend not only on the structure of an organic molecule, but also on the binding site features of the target protein. Thus, these new descriptors overcome the drawbacks of traditional molecular descriptors that depend only on the structures of organic molecules, regardless of what the target protein is. The new descriptors will be employed in conjunction with QSAR (quantitative structure activity relationship) modeling workflow to develop predictive models for selected protein targets (Aim-2a). Specifically, the two new methods developed in Aim-1 will be applied to build predictive models for targets from phosphodiesterase (PDE) and histone deacetylase (HDAC) families: PDE- 4, PDE-5, HDAC-7 and HDAC-8. We will then deploy the validated models via a web portal (Aim-2b) to benefit the research community of orphan neurodegenerative diseases. We aim to develop and apply innovative computational tools to study the protein targets of orphan neurodegenerative diseases, and to establish an open-access informatics resource to support the drug discovery efforts in these disease areas. We will ultimately contribute to alleviating the pain of orphan neurodegenerative disease patients.