How to choose aligners?

[hajime]

Dear all:

For some reasons, I need align my short read sequence as following conditions:

refseq: ATCCGATTGCCTCCAAATGCCCTAAATCGTA
my_sq: ATCC-AT-GCCTC-AAATGCCC-AAA-CG-A

(1) for the first 18 nt (red colored) from 5':
a. set as a seed
b. allow 3 mismatches (shown as red "-")
(2) Allow total 6 mismatches (rest mismatches as blue "-")

I tried to find some aligners, but I cannot find a good one.
For example, if I use bowtie with -n and allow 3 mismatch on seed, i cannot set the parameters for total 6 mismatches allowed.

Do you have any suggestions? Thanks a lot!!

Best,
Yi

[kga1978]

Give Mosaik a try - adjust -mm and -gop/-gap.

[tonybolger]

Dear all:

For some reasons, I need align my short read sequence as following conditions:

refseq: ATCCGATTGCCTCCAAATGCCCTAAATCGTA
my_sq: ATCC-AT-GCCTC-AAATGCCC-AAA-CG-A

Are these deletes or mismatches? If they are deletes, you can't use bowtie - it does substitution mismatches only, at least for version 1.

BWA may do what you want, but generally NGS aligners don't handle reads with many mismatches / indels all that well.

[hajime]

@Kga1978: Thanks for your suggestion. I'll try it.

@Tonybolger: the "-" indicates mismatches only (not small indel or other variant types). I'm sorry to let you feel confused. By the way, I'm not sure how the BWA can do what I want. Could you tell me more detail about that? thanks a lot

[tonybolger]

@Kga1978: Thanks for your suggestion. I'll try it.

@Tonybolger: the "-" indicates mismatches only (not small indel or other variant types). I'm sorry to let you feel confused.

OK, but normally '-' is used to as a gap-filler when there's a delete in a sequence.

By the way, I'm not sure how the BWA can do what I want. Could you tell me more detail about that? thanks a lot

A quick look at the BWA manual suggests that the combination "-n 6 -l 18 -k 3" will probably do what i think you want (you might also need/want to disable gaps).

[hajime]

OK, but normally '-' is used to as a gap-filler when there's a delete in a sequence.


A quick look at the BWA manual suggests that the combination "-n 6 -l 18 -k 3" will probably do what i think you want (you might also need/want to disable gaps).

Thanks for your kind and quick reply.

Actually, I did read BWA manual before I reply my last post.
However, I think I misunderstood and got confused about the meaning of "-n" parameter.

According to the description on the BWA website:
-------------------------
-n NUM Maximum edit distance if the value is INT, or the fraction of missing alignments given 2% uniform base error rate if FLOAT. In the latter case, the maximum edit distance is automatically chosen for different read lengths.
-------------------------

Based on your reply, I'd like to know whether you consider that "-n 6" and no gap allowed is the same as 6 mismatches in the read.

Thanks again!

[tonybolger]

Based on your reply, I'd like to know whether you consider that "-n 6" and no gap allowed is the same as 6 mismatches in the read.

I guess that it is, but i would suggest you test it to make sure it does what you want.

[HESmith]

I'm not certain if BFAST would work with such a small seed (you would need to modify the indexes), but it's very tolerant of mismatches.

[hajime]

Because I'm waiting for sequencing done, I cannot test any suggestions using real data right now. I'll tried to generate some fake reads or download others' data for furthered test.

Thanks for all the suggestions. Thank you guys!!

[marcowanger]

take a look at http://seqanswers.com/forums/showthread.php?t=15200