[feat] Generate VCF/BCF files for each alignment

[fanninpm]

My supervisor wants to know if there's an efficient way to generate VCF/BCF files using the outputs given in this workflow.

[DOH-JDJ0303]

Hi @fanninpm, is your goal to call variants? If so, vaper does not currently call variants, but this is something that could be added in the future. In the meantime, a VCF / BCF could be created from the read alignment and reference files that are saved for each sample:
(link to docs)

`${outdir}` 
├── `${sample}`
│   ├── align
│   │   ├── `${sample}`.bam
│   │   ├── `${sample}`.bam.bai
│   │   ├── refs.fa
│   │   └── refs.fa.fai

[DOH-JDJ0303]

It may be necessary to re-align the reads to a different reference genome, depending on what your goal is. In this case, VAPER outputs all the reads associated with each assembly:
(link to docs)

├── reads
│   ├── `${sample}`-`${reference}`_R1.fastq.gz
│   ├── `${sample}`-`${reference}`_R2.fastq.gz

[fanninpm]

After some tinkering, I think I came up with this patchwork-monstrosity of a workaround for now (relies on BCFtools):

OUTDIR=...
MPILEUP_DEPTH=$(nextflow config DOH-JDJ0303/vaper -o flat | grep 'params.cons_max_depth' | cut -f 2 -d '=' | sed 's/^[[:space:]]*//')

mkdir "${OUTDIR}/variants"
for sample_name in $(cut -f 1 -d ',' ${OUTDIR}/VAPER-summary.csv | tail -n +2 | sort); do
    [ -f ${OUTDIR}/${sample_name}/align/refs.fa.fai ] || continue
    for ref_name in $(cut -f 1 ${OUTDIR}/${sample_name}/align/refs.fa.fai); do
        bcftools mpileup --count-orphans --max-depth ${MPILEUP_DEPTH} --no-BAQ --min-BQ 0 --fasta-ref ${OUTDIR}/${sample_name}/align/refs.fa --regions ${ref_name} ${OUTDIR}/${sample_name}/align/${sample_name}.bam | \
            bcftools call --multiallelic-caller --variants-only --output-type v --output ${OUTDIR}/variants/${sample_name}_${ref_name}.vcf
    done
done

[DOH-JDJ0303]

Are you able to share what you plan to do with the variant data? This approach may not be strict / sensitive enough, depending on your goal.