I have a hard time to figure it out how correctly calculate the Alternative Allele Frequency (AAF) from Samtools mpileup.

I received the code which using samtools 1.8 mpileup (following code) get the following format:

.... samtools mpileup \
    -l ... capture_targets.bed  \
    -t DP,AD,ADF,ADR,SP,INFO/AD,INFO/ADF,INFO/ADR \
    -d100000000 \
    --output-BP \
    --output-MQ \
    --output-QNAME \ ...

Later in the code it, samtools mpileup data were converted to long format, clean up and add AAF As can be seen in the following code, the "DP4"which is " Number of high-quality ref-forward, ref-reverse, alt-forward and alt-reverse bases, FORMAT"was divided into 4 parts of "AD1", "AD2", "AD3", "AD4".

finally used the AD2/DP to calculate the AAF?

 ... %>% separate(AD,   c("AD1", "AD2", "AD3", "AD4")) %>% 
             mutate(AAF = as.numeric( **AD2**)/as.numeric(DP))

However, to me "**AD2**" is the "ref-reverse" and I think it should be "AD3" as "Alt-forward" for calculation of the AAF?

... %>% mutate(AAF = as.numeric( AD3 )/as.numeric(DP))

`

I think following code should be correct?

Can someone help me please!

Thanks a lot!