alignement and Variant calling with BWA and samtools
I am trying a variant calling pipeline with BWA and Samtools. Somehow when I have more sequences to align to references the output is less variant and I lose the Indels calls (even though the sequences carrying them are still present in the extended list). Could it be due to default parameters only calling variants if they are represented by a minimum % of the population?
Here is my pipeline:
bwa index ref_file
bwa mem ref_file all_genomes_file > alignement_file.sam
SAM to BAM:
samtools view -S -b s alignement_file.sam > alignement_file.bam
samtools sort alignement_file.bam -o sorted.bam
bcftools mpileup -B -f sorted.bam | bcftools call -mv --multiallelic-caller --variants-only > var.raw.vcf
• 92 views
The more reads you have, the more reliable the variant calling process is. If using more reads produces less variants, it means that those previously detected variants are not real variants.
From your pipeline I see that you're not removing or at least marking duplicates before calling variants. If your data comes from an hibridization enrichment process this could be a major issue you should address using duplicate removing tools such as Picard, sambamba, samblaster,...