Theoretical and Computational Approaches

该计划项目的总体目标是能够预测分子功能，例如， 烯醇化酶和AH中酶的底物专一性和/或专一性化学反应 超级家庭。计算项目的作用是整合和应用生物信息学 序列、结构和功能的表征，蛋白质的比较模型 结构，以及配体对接，以帮助实现这一目标。在与实验研究人员的密切合作中，我们设想了一个迭代周期，提供多个并行和串行路径，以获得对功能预测有用的高质量信息。将我们的方法应用于以前在结构和机械上描述的酶，将作为评估计算结果的控制。未知功能的建模序列将有助于为实验结构表征和生化测试选择靶点；相反，利用底物文库进行实验活性筛选的结果将用于改进超家族序列和结构的聚类，并为建模和对接练习提供额外的约束。我们最有希望的预测所针对的配体结构的实验解决方案将对于对接结果和方法的评估具有无价的价值。相反，环路建模练习可能有助于解释X射线结晶学无法解析的结构区域。我们预计，实验小组和计算小组之间的这种合作也将导致改进的工具和方法，用于半自动预测分子功能，特别是对于烯醇化酶和AH超家族，以及来自基因组计划的更广泛的未知或未充分描述的开放阅读框架(ORF)。

The overall goal of this Program Project is to enable prediction of the molecular functions, e.g., substrate specificity and/or specific chemical reaction, of enzymes in the enolase and AH superfamilies. The role of the computational project is to integrate and apply bioinformatic characterization of sequences, structures, and functions, comparative modeling of protein structures, and ligand docking to help achieve this goal. In close collaboration with the experimental investigators, we envision an iterative cycle providing multiple parallel and serial paths to obtaining high quality information useful for functional prediction. Applying our approaches to enzymes previously characterized structurally and mechanistically will serve as controls for evaluating computational results. Modeling sequences of unknown function will aid in the selection of targets for experimental structural characterization and biochemical testing; conversely, results from experimental activity screening with libraries of substrates will be used to refine clustering of superfamily sequences and structures, and to provide additional restraints for modeling and docking exercises. Experimental solution of liganded structures targeted by the most promising of our predictions will be invaluable for the evaluation of docking results and methodologies. Conversely, loop modeling exercises may aid in interpreting regions of structures that cannot be resolved by x-ray crystallography. We expect that this collaboration between the experimental and computational groups will also result in improved tools and methodologies for semi-automated prediction of molecular function, specifically for the enolase and AH superfamilies, and generally for the wider set of unknown or under-characterized open reading frames (ORFs) coming out of the genome projects.