Sniffles icon indicating copy to clipboard operation
Sniffles copied to clipboard

Common parameters still relevant in multi-sample mode

Open eliavw opened this issue 2 years ago • 3 comments

Hi,

From the --help page I've deduced that Sniffles categorizes its parameters in a few categories

  • Common parameters
  • SV Filtering parameters
  • SV Clustering parameters
  • Multi-Sample Calling / Combine parameters
  • SV Postprocessing, QC and output parameters

Based on this, one would expect that in multi-sample mode, the following categories are relevant

  • Common parameters
  • Multi-Sample Calling / Combine parameters
  • SV Postprocessing, QC and output parameters

That being said, I was wondering about some parameters in the first category, specifically:

  • reference
  • tandem-repeats
  • non-germline
  • phase

all of these are obviously relevant when calling variants, but to which extent do they still matter in multi-sample mode? It is my understanding that multi-sample mode focuses on combining already called SVs present in the individual snf files provided as inputs. It is unclear to me how e.g. a reference sequence would still matter in that case.

Thanks!

eliavw avatar Jul 07 '22 15:07 eliavw

Hi, I am not sure I am following. There is a section "Multi-Sample Calling / Combine parameters:" which lists the merging relevant parameters. The parameters are otherwise categorized for different steps in the program.

Thanks

fritzsedlazeck avatar Jul 07 '22 15:07 fritzsedlazeck

Hi,

Sorry for the confusion.

I understand that there is a section for parameters related to multi-sample calling, that is clear.

My question is about the "common parameters". Some of those evidently still apply in multi-sample mode (e.g. -i to specify inputs, -v to specify vcf file etc), whereas others like (--reference) seem only relevant when using Sniffles to call SVs in a single-sample.

Is that understanding correct?

Thank you!

eliavw avatar Jul 07 '22 15:07 eliavw

Maybe an example to make it more concrete;

sniffles --input sample1.snf sample2.snf ... sampleN.snf --vcf multisample.vcf --reference some_genome.fasta

In the command above, I would assume that specifying a reference (--reference some_genome.fasta) is irrelevant since the .snf files already contain called variants from prior (single-sample) runs of sniffles.

eliavw avatar Jul 07 '22 16:07 eliavw