While using SAMtools, consensus/variant calling used to give quality scores (eg. Consensus quality, SNP quality, RMS mapping quality). Can anyone suggest the computational logic behind these quality scores?
I am fairly sure these are Phred scores (Google knows about these), and give an indication of statistical robustness of the call. I think typically the original alignment algorithm probably generates these scores, and samtoolsl just reports them, but it depends what you are using samtools to do.