bcftools mpileup --max depth : What would be a reasonable setting for this option?

Question

[VBer]

Hello! I am trying to extract the consensus sequence from a sorted BAM file that was got after alignment. The commands I will be using are:

bcftools mpileup -Ou -f Aedes-aegypti-LVP_AGWG_CHROMOSOMES_AaegL5.fa AEB21A.bam | bcftools call -mv -Ob -o AEB21A_calls.bcf

bcftools index AEB21A_calls.bcf

cat Aedes-aegypti-LVP_AGWG_CHROMOSOMES_AaegL5.fa | bcftools consensus AEB21A_calls.bcf > AEB21A_consensus.fa

bcftools mpileup has a default of 250 for the --max-depth option, which I want to change. I am not sure what would be a good value to set. I am working with a eukaryotic insect genome of around 1.28 Gb.

[swbarnes2]

If you don't know how to set it, why not leave it at the default?

[VBer]

The default is 250 and the manual suggests to increase it.

Check the bcftools mpileup --max-depth option, most likely it should be increased. Note that by default only 250 reads per-file are considered at a position!

The previous samtools mpileup default was 8000. Hence I wanted to know what would be an appropriate value and how it is determined.

Answer 1

[Kevin Blighe]

Max read depth is not an area in which much work has been performed. The general [erroneous[ consensus is that 'more is better'. If you expect variants to be heterozygous or homozygous, then 30 read depth is sufficient. If you are expecting variants at low frequencies, e.g., somatic variants, then higher read depth may indeed be favourable.

[VBer]

Thank you for your reply. I had to run the command urgently and set the parameter to 8000 - the samtools mpileup default.