SAMTOOLS SNP calling question

[harrb]

Hi,

I have used tophat to map my paired-end Solexa reads to the mouse genome. I have called the SNPs using samtools and below is an exert from the raw outfile.
I wonder about the fact that for example for the second SNP the reference base is A the consensus base is R (i.e. a G or A) but the coverage is only 1 read. How can a single read have an R at this position? Anybody having any idea??


chr1 4336549 T K 9 9 60 2 ,g hK
chr1 4513199 A R 0 4 60 1 g C
chr1 4582311 G R 24 24 60 4 aa,, CChh
chr1 4583817 G A 6 125 60 3 AAa XXX
chr1 4588240 G T 3 5 60 2 TT 9/
chr1 4761834 G C 0 10 60 1 c I
chr1 4772806 T Y 21 21 60 5 ,$,,.C Ehhhh
chr1 4798537 T K 16 16 60 4 G.., Xhhh
chr1 4798553 A R 21 21 60 5 G$.,.. hHhhh
chr1 4798564 C M 18 18 60 6 ,...,^~a hhhhhh

[drio]

By default it calls everything even an heterozygous snp when you have only 1 read covering the position. You have to filter out the output (samtools.pl). One of the types of filtering is coverage.
Take a look to this.

[harrb]

thanks. I did this but just could not understand how a site can be called heterozygous when it is only represented by one read.