Alignment with BWA

[m4merg]

Hello!

Im using this BWA to align some reads

bwa index -p my_reference my_reference.fasta
bwa mem -t 4 my_reference input.fastq > output.sam 2

After what i found that in the position im interesteed in there are some bugs, or something like this. As shown on the picture i've attached in some reads motif TACCCGAGG in mutant reads algned with 3M1D6M with 1 mismatch, but in others as 9M with 1 mismatch. How can i make BWA to align all such reads as 9M?

[Brian Bushnell]

Perhaps I'm confused, but it looks like there are 3 classes of reads:

TACTCAAAGG (normal)
TACCCGAAGG (mutant 1)
TACCCGAGG (mutant 2)

The second class of mutant has a deletion in it, so of course it has to be aligned with a deletion somewhere.

[m4merg]

Perhaps I'm confused, but it looks like there are 3 classes of reads:

TACTCAAAGG (normal)
TACCCGAAGG (mutant 1)
TACCCGAGG (mutant 2)

The second class of mutant has a deletion in it, so of course it has to be aligned with a deletion somewhere.

So is there any way to align (mutant 2) as 6M1D3M with 2 mismatches rather than 3M1D6M with 2 mismatches, or I have to edit Sam file manually after each alignment? Because real mutations are T4C and A6G, but frequant sequance mistake 6delA in mutant reads disturb alignment and pileup consequantly and makes score of real mutations lower in SNP calling.

[Brian Bushnell]

You will have to manually edit the cigar strings. It's possible that if you align to the reverse-complement of the reference, the deletion will be put in the alternate location, though.