How samtools calculate error rate

Question

[jianzheng934963534]

I have aligned the fastq file to the reference genome and got a sam file. I know use "samtools stats" can get "error rate" of the sam file.

Is there any document shows how the error rate is calculated from the sam file? (since I want to seperately calculate the error rate of every read in the sam file instead of the average of the whole file)

I am really confused how they deal with softclips and hardclips when calculating the error rate.

Any answer is appreciated.

Answer 1

[Devon Ryan]

samtools adds up the number of mismatches (from the NM auxiliary tag) and divides that by the number of aligned bases. Soft and hard-clipped bases wouldn't be included, since they aren't aligned.