I can provide an example privately to demonstrate this if helpful, but basically if you have mutations within the same codon the resultant codon should be translated as the alternate AA rather than both of the SNPs being called separately as (incorrect) AA annotations.
Preferred behaviour would be: upon every annotation call the script would check to see if any other SNPs are present within the same codon, if yes, reconstruct the codon based on the ref then call those based on the reconstructed codon. It would also probably be a good idea to have an annotation field which states whether a codon has been reconstructed in this manner.
I can provide an example privately to demonstrate this if helpful, but basically if you have mutations within the same codon the resultant codon should be translated as the alternate AA rather than both of the SNPs being called separately as (incorrect) AA annotations.
Preferred behaviour would be: upon every annotation call the script would check to see if any other SNPs are present within the same codon, if yes, reconstruct the codon based on the ref then call those based on the reconstructed codon. It would also probably be a good idea to have an annotation field which states whether a codon has been reconstructed in this manner.