Lecture I Sequence Comparison

1
Lecture I Sequence Comparison
Bioinformatics Where Mathematics Meets Molecules

• MAA North Central Section Summer Seminar

2
Dot Plot A Visual Alignment
3
Dot Plot Sliding Windows
4
DotPlots on the Web
DOTLET http//www.isrec.isb-sib.ch/java/dotlet/Dot
let.html DNA DOT http//arbl.cvmbs.colostate.edu/
molkit/dnadot/ DOTTER http//www.cgb.ki.se/cgb/gr
oups/sonnhammer/Dotter.html
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
Edit Distance

• Linguistically constrained edit distance 5

9
A Linguistic Tree
10
Alignment With Gaps
TREE- -REED
TREE REED
TRE-E -REED
T-REE REE-D
Indel Insertion or Deletion Biologists
term for Gap
11
Global Sequence Alignment

• Use gaps to help align matching characters
• Compute similarity between words A and B

12
Alignment Scoring
TREE- -REED
TREE REED
1 3?
3 2?
TRE-E -REED
T-REE REE-D
3?? 2?
2 ?? 2?
13
Alignment Paths
A
C
G
G
C
T
C
Alignment of two sequences
A
T
G
Path through matrix
G
C
C
A-TGGCCTC ACGG-C-TC
T
T
14
Dynamic Programming

• Bellmans principle (i0,j0) (im,jn)
  (i0,j0) (i,j) ? (i,j) (im,jn)
• Recursive computation of cost function
  (Needleman-Wunsch)

Di,j max Di-1,j d(Ai, ), Di,j-1 d(,Bj),
Di-1,j-1 d(Ai,Bj)
i-1
i
j
j-1
15
Example
0
Initialization D1,j ? j Di,1 ? i
? -3 ? -4

ATGGCCTC ACGGC-TC
Optimal alignment
16
Local Sequence Alignment
Modified Recursion (Smith-Waterman)
Initialization D1,j ? Di,1 ?
0
Di,j max Di-1,j d(Ai, ),
Di,j-1 d(,Bj), Di-1,j-1
d(Ai,Bj), 0
17
Scoring Protein Alignments
18
Multiple Sequence Alignment
Image from http//bioinformatics.ubc.ca/resources/
tutorials/?ubicdocument_id26

• Exact higher dimensional DP
• Heuristic greedy algorithm (progressive)

19
BLAST

• Basic Local Alignment Search Tools
• Local alignments of query sequence with entire
  NCBI database
• Nucleotide, protein combinations
• Heuristic algorithm
• High-Scoring Segment Pairs (HSP)
• Word size parameter

20
Matching Sequence Summary
21
Alignment Scores
22
BLASTn Alignment
23
BLASTp and bl2seq
24
BLAST Statistics

• http//www.ncbi.nlm.nih.gov/BLAST/tutorial/Altschu
  l-1.html
• Extreme Value Distribution
• Raw Score --gt Bit Score --gt
  E-value