PROTEIN STRUCTURE AND EVOLUTION--A COMPUTER-BASED STUDY

正在寻求资金继续进行基于计算机的蛋白质研究 进化 主要目标是建立进化的根源， 现代蛋白质 为此，新出现的蛋白质 将序列输入计算机并进行搜索 与最新的序列数据库进行比对， 序列的候选人匹配受到适当的 统计措施，以评估相似性的可能性， 不是因为偶然。 匹配的序列被聚类到家族中 和进化树的构建。 在过去几年中 发现了大量意想不到的关系， 我们和许多其他人，在这一点上，很明显， 蛋白质原型的数量是一个可管理的数字。 即使有 银行里只有5,000个蛋白质序列，即使有很高的 数据库中的冗余度（即，相同的蛋白质 从许多物种），现在的机会甚至比这更好 来自哺乳动物的任何新确定的序列将 被发现类似于集合中已经存在的序列。 在这 考虑到，最后40个序列进入我们的收藏，更多 超过一半类似于已经报道的蛋白质，物种 除了冗余 这意味着我们已经处于一个良好的 一个位置，用于对序列进行分层分类， 他们的起源 我们特别感兴趣的是 脊椎动物血浆蛋白，其中许多是由于 一定量的“外显子改组”“这是一种现象， 对序列比较器提出了某些技术问题， 互换片段的长度通常只有40-50 残留物长。 该提案认为， 用于识别改组的外显子的程序，以及其他 旨在纠正结构性偏差的方法， 偶尔会使不相关的序列看起来相似。 需要 简单的程序，可以识别之间的关系， 蛋白质仅基于其氨基酸序列将 越来越多地被认为是大规模的测序项目（ 例如，完整的人类基因组）。

Funds are sought to continue a computer-based study of protein evolution. The major goal is to establish the evolutionary roots of contemporary proteins. To this end, newly appearing protein sequences are entered into a computer and searches conducted against up-to-date sequence data bases for potentially related sequences. Candidate matches are subjected to appropriate statistical measures to assess the likelihood that similarities are not due to chance. Matched sequences are clustered into families and evolutionary trees constructed. During the past several years a large number of unexpected relationships have been uncovered, by us and many others, and at this point it is clear that the number of protein prototypes is a manageable number. Even with only 5,000 protein sequences in the banks, and even with a high degree of reduncance in the data bases (i.e., the same proteins from many species), the chances are now better than even that any newly determined sequences from a mammalian organism will be found to resemble a sequence already in the collection. In this regard, of the last 40 sequences entered into our collection, more than half resemble already reported proteins, species redundancies aside. This implies that we are already in a good position for classifying sequences hierarchically with regard to their origins. We are especially interested in categorizing vertebrate blood plasma proteins, many of which are the result of a certain amount of "exon shuffling." This is a phenomenon that presents certain technical problems to the sequence-comparer, in that the lengths of interchanged segments are often only 40-50 residues long. The proposal considers the need for simple procedures for recognizing shuffled exons, as well as other approaches designed to correct for structural biases that occasionally make unrelated sequences appear similar. The need for simple procedures that can recognize relationships among proteins on the basis of their amino acid sequences alone will increasingly be felt as large-scale sequencing projects (the complete human genome, for example) are undertaken.