Affine gap penalty in Megablast

Question

[nazaninhoseinkhan]

Dear all,

I have a question. In megablast I don't understand why the existence and extension gap penalties in affine model are in that way. I thought the existence gap score must be always larger than the extension, but this is not true in megablast.

I will appreciate any help

Nazanin

Answer 1

[Istvan Albert]

The logic behind it is to allow single base insertions/deletions but penalize extending these. Megablast is tuned to find short, almost exact matches.

There are no rules like that one your expect for scores.

The only generic scoring matrix rule I heard of is that the scores of a scoring matrix should be mostly negative, and the negative values should be larger or comparable to positive ones otherwise a few matches can produce and maintain a runaway alignment.