ALGORITHMS FOR MACROMOLECULAR STRUCTURE ANALYSIS

大分子结构分析算法

• 批准号: 3374217
• 负责人: H LEE WEITH
• 金额: $ 12.84万
• 依托单位: PURDUE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 1989
• 资助国家: 美国
• 起止时间: 1989-09-01 至 1992-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3374217
• 关键词: computer simulation; mathematical model; model design /development; nucleic acid sequence; protein sequence

New algorithms will be developed to handle various analyses of nucleic acid and protein sequences. The algorithms will be designed to take advantage of massively parallel processors for high speed access and computations that will be required for the large amount of information expected to accumulate in the macromolecular sequence databases (especially as the human genome sequencing project gets under way). The algorithms will be used to conduct searching, pattern recognition, and alignments of sequences. They will then be packaged into user-friendly software for DEC and SUN workstations serving as front ends to the parallel processors. In addition, further development of Purdue University's secondary structure program for single-stranded nucleic acids will be undertaken. For this, new algorithms exploiting the simulated annealing approach will be designed to circumvent the problem of local energy minima encountered in this type of analysis.

将开发新的算法来处理核酸的各种分析 和蛋白质序列。这些算法将被设计成利用 的大规模并行处理器 大型数据库所需的高速访问和计算 预计在大分子序列中积累的信息量 数据库(特别是在人类基因组测序项目开始实施的情况下 方式)。这些算法将被用来进行搜索、模式 序列的识别和比对。然后它们将被打包成 作为前端的DEC和SUN工作站的用户友好软件 到并行处理机。此外，普渡大学的进一步发展 大学的单链核酸二级结构计划 将会被承担。为此，利用模拟的 将设计的退火法绕过局部最优解问题 在这种类型的分析中遇到的能量最小值。