MAQ and quality scores during alignment

[warrenemmett]

Hi,

I have seen some recent papers using MAQ and assume it is because it includes quality information?

Do you know of any better open source software that does this? I initially thought that quality scores weren't so useful, do they generally improve alignments or are they only useful in specific circumstances?

Thanks in advance!

[drio]

Hi,

I have seen some recent papers using MAQ and assume it is because it includes
quality information?

I guess you mean aligners that use base quality information.
What papers are you refering too? Did they recalibrate the quality scores?

That could be a good reason since it has proven to increase accuracy. Check the
latest alignment review from Homer and Li (Heng and Nils). In Fig 3 they show
how both novoalign (not opensource) and MAQ behave with and without using
base quality.

Do you know of any better open source software that does this? I initially
thought that quality scores weren't so useful, do they generally improve
alignments or are they only useful in specific circumstances?

With recalibration they seem to improve the accuracy of the alignments.
Bowtie also uses base quality. Novoalign is not opensource but seems to be
the most accurate aligner out there. So you may want to give it a try. You
can try it out for free.

[warrenemmett]

Thank you for the response!

I have heard of Novoalign and its reputation but I was hoping that there was a newer, open source algorithm available to replace MAQ, mostly because I want to use it and its a pain to implement it with large data volumes.

[drio]

BWA is the natural replacement, but it does not use base qualities. That should not stop you from using it. You'll see it preforms very well. It comes from the author of MAQ.

You can also try bfast.