Closed yuanjianwen-gif closed 2 years ago
The program bcftools call
assumes diploid or haploid genotypes and is not intended for detection of somatic mutations / RNA editing. For that, I recommend running like this
bcftools mpileup -a AD ... | bcftools call -mA ...
ignoring GTs and postprocessing the output using a custom script based on AD depths. The mathematical model of call -mv
(described here http://samtools.github.io/bcftools/call-m.pdf) cannot detect mosaic variants well.
I am using bcftools to call A to G transition in mRNA with RNAseq data, instead of snp at genomics ,and notice that there are mutations annotated 0/0 genotype but with high edited frequency (like DP4_alt = 50 , DP4_total = 200 ). It is really confusing what matters for genotype likehood calculation. Or it is not reasonable to use bctools to call mutation in mRNA