ITR/AP: Computational Analysis of Proteins: From Structure to Sequence to Function

蛋白质组学是基因组革命的下一步，准确确定蛋白质的结构和功能是蛋白质组学的重点。传统的分子生物学知识发现遵循了与功能基因组信息流相对应的研究范式，即从基因序列到蛋白质结构再到功能。这一计算生物信息学建议利用蛋白质结构和序列数据库的信息内容，遵循本建议标题中阐述的研究范式，即从结构到序列再到功能预测。这项研究的重要假设是，由计算几何(Delaunay镶嵌)定义的空间最近邻残基基序为蛋白质结构和功能提供了独特的决定因素。该建议的主要目标包括：(1)基于蛋白质结晶学数据库的非冗余子集开发和验证多体统计势；(2)通过使用蒙特卡罗采样、统计势和特定残基基序定义的空间约束的链生长算法的新实现来预测蛋白质的三维结构；(3)确定已知蛋白质的序列和结构特定的氨基酸残基基序，以及大规模预测基因组序列的蛋白质结构和功能类。这一提议的成功实施将对功能基因组学和蛋白质组学产生广泛影响。

Accurate determination of protein structure and function is the focus of proteomics, the next step in the genomic revolution. Traditional knowledge discovery in molecular biology has followed a research paradigm corresponding to the information flow in functional genomics, that is from gene sequence to protein structure to function. This computational bioinformatics proposal exploits the information content of the protein structural and sequence databases following a research paradigm formulated in the title of this proposal,namely from structure to sequence to function prediction. The important postulate of this research is that spatial nearest neighbor residue motifs defined by the means of computational geometry (Delaunay tessellation) provide unique determinants of protein structure and function. The major goals of this proposal include: (1) development and validation of multibody statistical potentials on the basis of non-redundant subsets of the protein crystallographic database, (2) prediction of 3D protein structure by the means of novel implementation of the chain growth algorithm using Monte Carlo sampling, statistical potentials, and spatial constraints defined by specific residue motifs, and (3) determination of sequence and structure specific amino acid residue motifs for known proteins and large scale prediction of protein structural and functional classes for genomic sequences. The successful implementation of this proposal will have a broad impact on functional genomics and proteomics.