Proteolysis in silico: statistics, structural chemistry, and biology

描述(由申请人提供)：该提案将提供预测蛋白质分解事件的信息、新算法和计算工具。最终的目标是在蛋白质组范围内对任何给定的蛋白酶底物进行准确的预测。然而，我们目前的工作将主要集中在基质金属蛋白酶(MMPs)、半胱氨酸蛋白酶(Caspase)和属于丝氨酸蛋白酶家族的几种蛋白质转换酶(PC)上，因为桑福德-伯纳姆医学研究所(Sanford-Burnham Medical Research Institute)已经有大量关于这些蛋白酶的实验信息。我们的方法可以很容易地扩展到任何其他蛋白酶，当有大量统计意义的底物可用于推导特异性图谱时。提出的预测方法的独特之处在于将基于序列的预测与其他因素相结合。这些包括：底物的结构特征，协同作用，以及底物和蛋白酶的共定位和共表达。我们还将包括蛋白质底物中裂解位点附近残基的SNPs(单核苷酸多态)和PTMS(翻译后修饰)的信息。这两种作用可以通过关闭蛋白分解事件或通过创建新的可能的切割位点来修改蛋白分解事件。这样的改变可能会导致疾病或症状。蛋白降解事件，例如，蛋白酶-底物对，将被映射到已知的调控网络上。所有收集的信息和开发的工具将在PMAP网站(www.Proteolysis.org)上免费提供，供生物医学研究界使用。由于蛋白酶通常有十几种底物，而且这些底物在正常生理和病理上往往不同，这个项目的影响可能是巨大的。我们的预测不是在一个接一个的基础上识别蛋白酶底物，而是产生一组注释很好的底物，很可能具有生物学意义。

DESCRIPTION (provided by applicant): This proposal will provide information, new algorithms, and computational tools for predicting proteolytic events. The ultimate goal is to make accurate proteome-wide predictions of the substrates for any given protease. However, our current effort will focus mainly on matrix metalloproteases (MMPs), caspases, and several protein convertases (PCs) belonging to the serine protease family because a vast amount of experimental information on those proteases is already available at the Sanford-Burnham Medical Research Institute. Our approach can be easily extended to any other proteases when a statistically significant number of substrates become available for deriving a specificity profile. The unique feature of the proposed prediction method is combining sequence-based predictions with other factors. These include: structural features of the substrates, cooperative interactions, and co-localization and co-expression of substrates and proteases. We will also include information about SNPs (single nucleotide polymorphisms) and PTMs (posttranslational modifications) of the residues in the vicinity of the cleavage sites in protein substrates. These two effects can modify the proteolytic event by turning it off or by creating a new possible cleavage site. Such modifications can lead to diseases or syndromes. The proteolytic events, e.g., protease-substrate pairs, will be mapped onto the known regulatory networks. All the information that is collected and tools that are developed will be freely available on the PMAP Web site (www.proteolysis.org) for use by the biomedical research community. Because proteases usually have more than a dozen substrates, and because the substrates often differ in normal physiology vs. pathology, the impact of this project could be immense. Rather than identifying protease substrates on a one-by-one basis, our predictions will produce very-well-annotated sets of substrates that will likely have biological significance.