Yep, will do.
I'm trying to do everything with common bash utilities so that users won't have to install anything. I think my ugly fix is working:
## getting the line numbers of the bad sequences
grep -v '^>' aligned-dna-sequences.fasta | grep [^THADGRC\.SMBNWVKY-] -n > badseqs.txt
## reformatting these line numbers to delete
for val in $(cut -f 1 badseqs.txt -d ':'); do
myvar1=$(($val*2))
myvar2=$(($myvar1-1))
echo $myvar2"d"
echo $myvar1"d"
done | tr '\n' ';' > droplines.txt
## deleting these lines from original fasta
myval=$(cat droplines.txt)
sed -e $myval aligned-dna-sequences.fasta > cleaned_aligned-dna-sequences.fasta