Anders Sune Pedersen
Anders Sune Pedersen
@maxulysse @FriederikeHanssen @amasplund : I'll run `bcftools sort z` followed by `tabix -p vcf` on the concatenated vcf-file, ok?
So here I ran deepvariant, mpileup, strelka and freebayes, and I get the same variant called three times: ``` ##bcftools_concatCommand=concat --output testN.vcf.gz --threads 1 testN.deepvariant.vcf.gz testN.bcftools.vcf.gz testN.strelka.variants.vcf.gz testN.freebayes.vcf.gz; Date=Mon Nov...
> yeah, is the meta info also the same across all? Not sure I understand what you mean 🤔 The different variant-callers produces vcf-files with different INFO- and FORMAT-fields.
Here is the concatenated (and sorted) vcf-file resulting from the following cmd: ``` nextflow run main.nf -profile test,singularity --input mapped_joint_bam.fixed.csv -dump-channels -ansi-log false --step variant_calling --concatenate_vcfs --tools deepvariant,freebayes,strelka,mpileup ``` [testN.germline.vcf.gz](https://github.com/nf-core/sarek/files/10105274/testN.germline.vcf.gz)...
Just for the record: the cnvkit doesn't produce any vcf-files, so the concatenated vcf-file doesn't contain variants from the cnvkit.
> Some comments: > > * My experience is that CNVkit works really bad without a reference dataset as it uses coverage distributions (log2ratios I think) to determine if something...
Coming back to the point about adding an INFO-field to the vcf-files before concatenating them: @amasplund come up with a nice, little bash+awk-script for adding some "constant" INFO-field to a...
> > So far I am just concatenating the germline-vcfs from haplotypecaller and strelka, and placing the resulting vcf `.germline.vcf.gz` in the results-folder `results/variant_calling/concat/`. > > I think it's best...
@FriederikeHanssen @maxulysse : Can I get you guys to do a preliminary review of this PR? If this PR looks okay, then I'll update the corresponding modules in github.com/nf-core/modules. I've...