Identification of functionally important protein residues by means of entropy based methods, and experimental validation by mutational analysis

我们的第一代程序H2r通过基于熵的方法分析相关突变，以确定蛋白质中的关键残基位置。在目前的资助期内，通过对参与色氨酸生物合成的两种酶的突变分析，实验验证了H2r的预测。结果表明，氨基酸在鉴定位置的交换改变了催化效率或蛋白质的稳定性。在这些有希望的结果的基础上，我们希望进一步提高H2r的特异性和敏感性。用基于von Neumann熵的相关测度代替Shannon熵来提高特异性。这使我们能够考虑残基对的物理和化学性质。在当前的发现阶段准备了相应的相似性矩阵，并实现了第一个原型算法。如果H2r利用所有统计上显著的相关信号，灵敏度将增加。一个特定于msa的零模型和一个特定于残差对的计算截止值的方法即将完成，并将在未来几个月内集成。由此产生的算法将被用于识别单个蛋白质中比迄今可能的更大比例的关键位置。除了预测单个蛋白质的关键残基位置外，我们计划将相关分析扩展到蛋白质-蛋白质复合物。这些位置将在计算机上进行评估，预测将通过使用表征良好的酶和酶复合物的突变实验进行验证。

Our first generation program H2r analyses correlated mutations by means of entropy-based methods in order to identify crucial residue positions in proteins. In the current funding period, predictions made by H2r have been experimentally tested by mutational analysis of two enzymes involved in the biosynthesis of tryptophan. It turned out that amino acid exchanges at the identified positions alter catalytic efficiency or protein stability. Building up on these promising results, we want to further increase specificity and sensitivity of H2r. Specificity shall be improved by replacing Shannon’s entropy with correlation measures based on the von Neumann entropy. This allows us to consider physical and chemical properties of residue- pairs. Corresponding similarity matrices were prepared during the current finding period and a first prototypic algorithm was implemented. Sensitivity will increase, if H2r exploits all statistically significant correlation signals. An MSA-specific null-model and a residue-pair-specific approach to compute a cut-off are nearly finished and will be integrated within the next months. The resulting algorithms will be utilized to identify a larger fraction of crucial positions within individual proteins than hitherto possible. In addition to predict crucial residue-positions in individual proteins, we plan to extend correlation analysis to proteinprotein complexes. These positions will be assessed in silico and predictions will be validated by mutational experiments using well characterized enzymes and enzyme complexes.