Shall I change bcftools mpileup max depth?

Question

[marongiu.luigi]

Hello,

I would like to run a variant calling with bcftools; I am following the lead from this post and it is said to increase the coverage with --max-depth since it is by default 250.

Shall I really change it or it is due to a previous version? I am using bcftools 1.9-26-g73a628e-dirty

And what value should be the coverage to provide?

Thank you

[gb]

It depends on what you want yourself so maybe you should add that to your question. If you want to call variants and sequence really deep and need to be absolutely sure about certain variants you can put it on 999999 so to speak.

[marongiu.luigi]

well, of course, I would like to have variants with high probability.

[ATpoint]

You should ask yourself if these variants are reliable. If you have an average coverage of e.g. 30 and a variant supported by 10000 reads then odds are good that many are false-positive alignments due to low complexity regions. That is why this cutoff option exists.