Understanding the dpal logic

[karen-nativ]

Hello,
I'd like help understanding the logic of the dpal module in global mode.
It should find the best global alignment using the Needleman Wunsch algorithm correct?
However I see different behavior, for example when running the following (the m flag allows a gap of length 100, and g parameter for global mode):
./ntdpal -m 100 ACCCCTTTTTTTTTTTTT AT g

The output is:
|ACCCCTTTTTTTTTTTTT| |AT| G score=0.00 len=2 |16,0|17,1|

From the output I understand that the alignment created is:

ACCCCTTTTTTTTTTTTT
----------------AT

Meaning the alignment includes indels and at the end a T/A mismatch and a T/T match.
However the Needleman Wunsch algorithm in this case can give us the following alignment:

ACCCCTTTTTTTTTTTTT
A----------------T

In this case the created gap is of the same length but we have two matches - A/A and T/T.
So it seems this output has a better score.

Why are the outputs different in this case?
Or perhaps does the dpal module work a bit differently than the NW algorithm? If so, how?

Thanks for the help.

[untergasser]

I am not sure it is a complete implementation. It might be optimized for speed and primer design. I think match scores 1 mismatch and gap -1. So it the first is better. I hope this helps.

Best,
Andreas