mpileup SNP detection

[SWP]

Hello all-

I've got several full genomes sequenced off a HiSeq. There's good coverage (~30x). Alignments were performed with default parameters in BWA. These are Caucasian individuals.

Variant detection was performed using mpileup, using almost exact parameters listed on the samtools page:

samtools mpileup -uf ref.fa output.sorted.bam | bcftools view -bvcg - > var.raw.bcf

bcftools view var.raw.bcf | vcfutils.pl varFilter -D100000 > var.flt.vcf

Not a single indel was called, and only ~2 million SNPs. 2 million is well below the >3 million I expected.

I was wondering if anyone who has aligned a full genome with similar depth of coverage, and called variants using mpileup and got a more reasonable variant number, would be willing to share what parameters were used?

The raw bcf file has ~3.9 million variants.

Any ideas/suggestions?

[colindaven]

That does seem like a very high coverage you are requiring (-D100000).

Perhaps try a lower number? Mpileup is pretty good software but I agree with what you are saying.

[SWP]

The -D option is not a minimum coverage requirement, it's the highest coverage allowed...unless I completely misunderstand something about mpileup. My intention was to have no regions excluded because the coverage was too high, and since I'm a bit OCD I used 10000 as opposed to something like 300 which would probably be adequate.

I believe I had a corrupted file and I'm running mpileup now and it appears to be working correctly.