Guerrucci M A, Bellé R
Atelier de Bioinformatique, Institut Curie, Paris, France.
Biosci Rep. 1995 Jun;15(3):161-71. doi: 10.1007/BF01207456.
Using an approach for protein comparison by computer analysis based on signal treatment methods without previous alignment of the sequence, we have analysed the structure/function relationship of related proteins. The aim was to demonstrate that from a few members of related proteins, specific parameters can be obtained and used for the characterisation of newly sequenced proteins obtained by molecular biology techniques. The analysis was performed on protein kinases, which comprise the largest known family of proteins, and therefore allows valid estimations to be made. We show that using only a dozen defined proteins, the specific parameters extracted from their sequences classified the protein kinase family into two sub-groups: the protein serine/threonine kinases (PSKs) and the protein tyrosine kinases (PTKs). The analysis, largely involving computation, appears applicable to large scale data-bank analysis and prediction of protein functions.
我们采用了一种基于信号处理方法的计算机分析蛋白质比较方法，无需对序列进行预先比对，从而分析了相关蛋白质的结构/功能关系。目的是证明，从少数相关蛋白质成员中可以获得特定参数，并用于表征通过分子生物学技术新测序得到的蛋白质。分析针对蛋白激酶进行，蛋白激酶是已知最大的蛋白质家族，因此能够进行有效的评估。我们表明，仅使用十几种定义明确的蛋白质，从其序列中提取的特定参数就能将蛋白激酶家族分为两个亚组：蛋白丝氨酸/苏氨酸激酶（PSK）和蛋白酪氨酸激酶（PTK）。该分析主要涉及计算，似乎适用于大规模数据库分析和蛋白质功能预测。
