BWA and gaps

[efoss]

In the BWA manual, I see options referring to "gap opens" and "gap extensions". What does this mean? Also, for a typical user who wants to identify SNPs, indels, etc. in a data set, is there are the default options with regard to gaps likely pretty optimal? I realize that you can always tweak stuff, but I started wondering more about messing around with gap options when I saw a paper refer to "gap-enabled BWA". This made me concerned that without messing around with these options, my BWA was not "gap-enabled", but as far as I can tell, the default options are good for dealing with gaps.

Thanks.

Eric

[rskr]

In the BWA manual, I see options referring to "gap opens" and "gap extensions". What does this mean? Also, for a typical user who wants to identify SNPs, indels, etc. in a data set, is there are the default options with regard to gaps likely pretty optimal? I realize that you can always tweak stuff, but I started wondering more about messing around with gap options when I saw a paper refer to "gap-enabled BWA". This made me concerned that without messing around with these options, my BWA was not "gap-enabled", but as far as I can tell, the default options are good for dealing with gaps.

Thanks.

Eric

It is enabled by default in BWA, but gap-enabled distinguishes it from similar pieces of software which aren't gap enabled, or even gap capable, which of course isn't a selling point so much as being gap enabled.

[efoss]

It is enabled by default in BWA, but gap-enabled distinguishes it from similar pieces of software which aren't gap enabled, or even gap capable, which of course isn't a selling point so much as being gap enabled.

Thanks. That makes sense.

And to the other part of my question: What are "gap opens" and "gap extensions"?

Thanks.

Eric

[nilshomer]

the penalty of of a contiguous indel of length L is scored by F(L) = O + L*E, where O is the gap open penalty and E is the gap penalty. The first gap in the indel is scored (O+E), while the following gap (extensions) have score E. I would recommend buying an intro bioinformatics book like Jones and Pevzner.

[efoss]

the penalty of of a contiguous indel of length L is scored by F(L) = O + L*E, where O is the gap open penalty and E is the gap penalty. The first gap in the indel is scored (O+E), while the following gap (extensions) have score E. I would recommend buying an intro bioinformatics book like Jones and Pevzner.

Thanks, Nils Homer. I just put in an order for your suggestion on Amazon.

Eric

[yfwang]

By the way, what is the difference with SW algorithm?