Global Alignment with Scoring Matrix and Affine Gap Penalty solved by 245

July 2, 2012, midnight by Rosalind Team

Topics: Alignment, Dynamic Programming

Mind the Gap

In “Global Alignment with Scoring Matrix”, we considered a linear gap penalty, in which each inserted/deleted symbol contributes the exact same amount to the calculation of alignment score. However, as we mentioned in “Global Alignment with Constant Gap Penalty”, a single large insertion/deletion (due to a rearrangement is then punished very strictly, and so we proposed a constant gap penalty.

Yet large insertions occur far more rarely than small insertions and deletions. As a result, a more practical method of penalizing gaps is to use a hybrid of these two types of penalties in which we charge one constant penalty for beginning a gap and another constant penalty for every additional symbol added or deleted.

Problem

An affine gap penalty is written as $a + b \cdot (L-1)$, where $L$ is the length of the gap, $a$ is a positive constant called the gap opening penalty, and $b$ is a positive constant called the gap extension penalty.

We can view the gap opening penalty as charging for the first gap symbol, and the gap extension penalty as charging for each subsequent symbol added to the gap.

For example, if $a = 11$ and $b = 1$, then a gap of length 1 would be penalized by 11 (for an average cost of 11 per gap symbol), whereas a gap of length 100 would have a score of 110 (for an average cost of 1.10 per gap symbol).

Consider the strings "PRTEINS" and "PRTWPSEIN". If we use the BLOSUM62 scoring matrix and an affine gap penalty with $a = 11$ and $b = 1$, then we obtain the following optimal alignment.

 PRT---EINS
 |||   |||
 PRTWPSEIN-

Matched symbols contribute a total of 32 to the calculation of the alignment's score, and the gaps cost 13 and 11 respectively, yielding a total score of 8.

Given: Two protein strings $s$ and $t$ in FASTA format (each of length at most 100 aa).

Return: The maximum alignment score between $s$ and $t$, followed by two augmented strings $s'$ and $t'$ representing an optimal alignment of $s$ and $t$. Use:

Sample Dataset

>Rosalind_49
PRTEINS
>Rosalind_47
PRTWPSEIN

Sample Output

8
PRT---EINS
PRTWPSEIN-

Please login to solve this problem.