《生物信息学：导论与方法》----序列对比----听课笔记（四）

第二章  序列对比

2.6 学生课堂报告1

• 序列对比有两个特性：分解为重叠子问题和优化之后能得到最小子结构，这两个正好符合动态规划的思路。
• Needleman-Wunsch算法的打分矩阵填充并不一定要从左上角到右下角，也可以从右下角到左上角，凭个人喜好。
• Whether a particular result differs significance from a fortuitous(偶然) match between two random sequencees?
• To answer this question we can do: 1. Sequence alignment between two sets of random sequences;两条序列比对的结果与这个随机得到的分布的结果具有显著性差异的话，就说明匹配得到的结果和随机得到的结果是不一样的。 2.Sequence alignment between one set of random sequences and a real sequence.
• 疑问：显著性差异如何比较？一条序列和一个分布比？
• Local alignment compared with global alignment

1. Zero could terminate the current local alignment
2. Mismatch must be negative scored

• Local alignment----other properties

1. Suitable to identify conserved local sequence(substring)
2. Guaranteed to find the best local alignment
3. Perform poorly when dealing with separated regions within a long sequence.(不太理解这一点)

2.7 学生课堂报告2

• 
• 
• How to choose appropriate scoring matrix?
• Scoring matrix-----DNA scoring matrix: Unitary matrix(等价矩阵)； Transition-transversion(转换-颠换矩阵)；BLAST matrix
• Protein scoring matrix: PAM(Point Accepted Matrix); BLOSUM(Block Substitution Matrix)
• 
• PAN矩阵表示替换的概率，BLOSUM表示的是两者的相似性。
• 
• Semi-global alignment
• Use case:When one of the sequences is significantly shorter than the other.
• 
• The penalty of gap is more than the penalty of mispairing.
• The latter the gap appears, the more penalty increase.
• 罚分规则是可以根据需求改变的。