What Is The Difference Between Samtools Mpileup And Pileup

Question

[Ian]

As of samtools 0.1.8 'mpileup' appears to have superseded 'pileup' for variant detection. The samtools website now has pileup as deprecated.

So i am wondering whether i should repeat my recent SNP/INDEL calling with mpileup? But does anyone know what the improvements are using mpileup?

For those who don't already know 'pileup' and 'mpileup' can be used to create a consensus sequence using NGS mapped reads relative to a reference genome, but can also highlight variants relative to the reference.

Answer 1

[lh3]

For single-sample SNP calling, they differ little. Pileup uses BAQ by default, too. The major difference comes from indel calling. Mpileup implements a more advanced method. For indel calling, do not use pileup any more.

[lh3]

Mpileup also outputs pileup.

[Drio]

It seems pileup is going away in the next release. The pileup raw output was still very useful (read coverage among others). If that feature goes away, what tool can be used as a replacement? An alternative would be to write your version. I wanted to check before moving in that direction.

Answer 2

[Pablo]

Indeed, mpileup is a pileup that supports multiple samples and the old pileup is deprecated.

Mpileup by default uses a new algorithm (BAQ). The Snp calling is now done by BcfTools, which (I think) also made some improvements.

Since generating mpileup/bcf/vcf is not such a long process, I'd recommend re-calculating the calls.