序列两两比对算法_序列比对算法详解.ppt

PsiBlast介绍 Psi (Position Specific Iterated) 是一个自动的 profile-like搜索 程序首先执行一个有空位的对数据库的Blast搜索。然后，重要的比对信息被用于构建一个“特定位置”打分矩阵。 该矩阵在下一轮数据库搜索中替换了查询序列 程序将迭代不断运行，直到发现新的有意义的信息 * 思考题 问题1：为什么说如果在一个比对中获得的P-值小(<0.001) ，则比对可能具有生物学意义？P值概念：一个变量取值严格说碰巧大于等于观察值的概率记为P(z>zo) 问题2 ：如何理解“成对序列的比对是窃窃私语… 多序列比对是大声嚷嚷” ——Hubbard et al 1996 * * * What you want to learn when evaluating an alignment is whether it is random or meaningful. If it is meaningful, then how much? * * * * All pairs of sequences are aligned separately in order to calculate a distance matrix indicating the divergence of each pair of sequences. ? A guide tree is calculated from the distance matrix. * ? Once a gap, always a gap. The positions of the gaps that were introduced during the early alignments of closely related sequences are not changed as new sequences are added. (The placements of gaps in alignments between closely related sequences are much more accurate than those between distantly related ones.) * ? The final multiple alignment is performed by a series of progressive, pairwise alignments between sequences and clusters of sequences, according to the branching order in the guide tree. ? Since the alignment is calculated on a progressive basis, the choice of sequences in the initial stage of alignment can affect the final alignment. In addition, anything that affects the calculation of the guide tree such as different similarity matrices or gap weights can also affect the multiple alignment. * Averaged scores used at each aligned position A A C . . . 0 -1 -2 -3 . . . A -1 T -2 C -3 . . . . . . 1 0 -1 0 1 0 -1 0 2 C C AC TC AAC ATC 追朔 : 比对的构建 (5) * 数学的表达 0 0 2 0 1 0 ￡ < - = ￡ < - = l j j S l i i S j i 序列 1长度 序列 2长度 0 f o r 1 ) , ( w i t h 0 , 0 ) , ( m a x 2 1 1 , 1 , 1 , 1 ? í ì = = ￡ < ￡ < ? ? ? í ì - + - = - - - - e l s e b a b a s l j l i 1 S b a s S 1 S S j i j i j i j i j i j i i j 初始化 打分 * 动态规划计算的复杂度 计算时间: O(nm)，这里 n 和 m 是序列的长度 检索时间: O(Max (n,m)) [worst case: n+m; best case: Min(n,m)] 需要的内存