4

String alignment

instructor: Ross A. Lippert

http://www-math.mit.edu/~lippert/18.417/

For class:

modification to the schedule
we have a grader/TA

 gacttaagactttat
a *   ** *    * 
c  *      *     
c  *      *     
a *   ** *    * 
a *   ** *    * 
g*      *       
g*      *       
g*      *       
t   **     *** *
t   **     *** *

Making change

Noticing a recurrence

A solution to the change problem can be expressed in terms ofsimpler solutions.

Making Change Recursively

In pseudocode

Just one problem: still O(M^d) !!!

A look at the Change search tree:

For 77 by (1,3,7)

Solution: store intermediate results

This is dynamic programming

Files of strings (FASTA format)

FASTA (fasta-A) format is ubiquitous

>sequence_id comments
atgc...
atgc...
...
>sequence_id comments
atgc...
atgc...
...
...

The initial lines are called deflines and sometimes get used to store all sorts of extra information.

Some programs impose or expect line limits on the sequence lines

Alignments

An example alignment of ATCTGATG and TGCATAC

matches
mismatches
in/dels

Global Alignment by Recurrence

I found a cute demo at this site

Variation 1: Longest common subsequence

Variation 2: Sparse alignments

The problem:

Input: (x₁,y₁,Score₁),(x₂,y₂,Score₂)... regions of high similarity.
Output: a subset where x_i < x_i+1 and y_i < y_i+1 with maximum score

Score = [Scores] - [Gaps]