asked like that you question cannot find an answer! if you have different populations with frequencies of alleles for a given SNP, the question might be: how to recognize the ancestral among the two (or more) alleles? By default one can say that the ancestral allele (according to neutral theory and a large panmictic population) is the one that has the highest frequency, since the frequency of new alleles increases very slowly.. however if data on many populations are available there might be evidence for selection or founder effect or genetic drift .. that can help in inferring precisely the ancestral allele (that carried by the TMRCA: the most recent common ancester)..