I get the same problem with mutations to thymine in Coot. Mutations to my DNA strands all end up as RNA (i.e. mutation of guanine to cytosine results in "Gd" to "Cr"). The problem is worse with thymine as I can't even make the mutation.
Also, mutation of thymine to other bases results in the retention of the "C7" methyl group, so that mutation of thymine to cytosine results in a base that looks a lot like thymine and mutation of thymine to adenine results in a base that looks like adenine with a floating C7 atom. I then have to go and delete the atom for Refmac to run correctly.
