My unreliable consensus sequences with mpileup?

[lg36]

Dear All,

I've used (what I think are) default settings to align paired end Illumina data using bwa then samtools and finishing with mpileup.

Before obtaining my sequences I PCR'd all our strains in a specific area of the genome to confirm their family according to the presence or absence of certain SNP's.

The problem is that the consensus fasta that is derived from mpileup always has the family of the reference strain and not the query strains that I've lined up (despite having good coverage in the region that defines the family).

Somehow the reference sequence is being output rather than the query sequence at this part of the genome.

Can anybody help?

Best wishes and thanks

[Heisman]

Why don't you just call SNPs and compare the SNP calls?

[lg36]

Thanks for your reply.

I was wanting to somehow calibrate the SNP filters based on the correctness of the consensus sequence. I figured (perhaps incorrectly) that a reasonable test of the consensus sequence (and hence the SNP calls) would be to check the consensus sequence against confirmed oligonucleotides in the lab. Does this make any sense.

Can it be that the SNP's called are totally different from the consensus sequence.