I notice on dbSNP rs6352 has an ancestral allele of G - http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=rs6352
The hapmap in humans for this allele is very rare, the homozygous version practically not existing at all http://hapmap.ncbi.nlm.nih.gov/cgi-perl/snp_details_phase3?name=rs6352&source=hapmap3r3_B36&tmpl=snp_details_phase3
My understanding has been that the ancestral allele refers to the common gene that an ancestor in all humans once shared. Now, I thought the ancestral allele was usually the most frequent in a population, but I can see this was most likely a wrong assumption.
Was it a wrong assumption, or was this example above just a rare exception? Is deviation from ancestral alleles nearly worthless as a metric to investigate possible mutated alleles that may be leading to disease? I realize the proper answer may be too long, but just a brief summary would be appreciated.