strange results of samtools

[liying]

Hi, everyone！
I'm dealing with 14 sets of 81*2 data. After mapping them to reference genome by bwa 0.5.9-r16, I try to call variants using samtools 0.1.18. The strange thing is that all the thousands results of each sample are INDEL, though I can see there are SNPs by IGV.
Do you have any idea about this?


my command:
samtools mpileup -6AB -Q 30 -uf Egrandis_162.fa 1.rmdup.bam | bcftools view -bvcg - >1_30.RawVar.bcf 2>mpileup.log&
bcftools view 1_30.RawVar.bcf | vcfutils.pl varFilter -Q 30 -d 2 - > 1_30_30.FilterVar.vcf

[swbarnes2]

Originally Posted by liying

Hi, everyone！
I'm dealing with 14 sets of 81*2 data. After mapping them to reference genome by bwa 0.5.9-r16, I try to call variants using samtools 0.1.18. The strange thing is that all the thousands results of each sample are INDEL, though I can see there are SNPs by IGV.
Do you have any idea about this?


my command:
samtools mpileup -6AB -Q 30 -uf Egrandis_162.fa 1.rmdup.bam | bcftools view -bvcg - >1_30.RawVar.bcf 2>mpileup.log&
bcftools view 1_30.RawVar.bcf | vcfutils.pl varFilter -Q 30 -d 2 - > 1_30_30.FilterVar.vcf

Start by finiding a SNP that you see in IGV, and looking at the pileup in that position, and then the sam file in that region. Do the reads have good quality scores at the locus? Is the mapping quality alright? Something that I've observed, though it might not explain all your missing SNPs, is that sometimes, the BAQ calculations will mess up real SNPs, but not indels. I've occasionally seen this with sanger verified SNPs. So try rerunning mpileup with -B, see if that makes a difference.

[liying]

Originally Posted by swbarnes2

Start by finiding a SNP that you see in IGV, and looking at the pileup in that position, and then the sam file in that region. Do the reads have good quality scores at the locus? Is the mapping quality alright? Something that I've observed, though it might not explain all your missing SNPs, is that sometimes, the BAQ calculations will mess up real SNPs, but not indels. I've occasionally seen this with sanger verified SNPs. So try rerunning mpileup with -B, see if that makes a difference.

Thanks for you reply.
I'm sure these SNPs should pass my fliter. The MAPQ, baseQ and depth is good enough. And no matter how low threshold be set, I did not got any SNPs anyway.
As -B option was already be used, I presumed you mean without it. That didn't help.

Perhaps I should give up and try other SNPcaller.

PS:GATK is really complicated. I'm trying.

[liying]

Foget it. I'm a fool. I shouldn't use -6 option after adding -I in bwa.
Well, it's a good thing that I'm wrong and samtools is right, so I can use it...